J. Cent. South Univ. (2017) 24: 360-373

DOI: 10.1007/s11171-017-3438-z

Evaluation of rutting performance of stone matrix asphalt mixtures containing warm mix additives

Rezvan Babagoli¹, Hassan Ziari²

1. Department of Civil Engineering, Faculty of Engineering, Pardis Branch, Islamic Azad University,

Pardis, Tehran, Iran;

2. School of Civil Engineering, Iran University of Science and Technology, Tehran

Central South University Press and Springer-Verlag Berlin Heidelberg 2017

Abstract: Permanent deformation or rutting, one of the most important distresses in flexible pavements, has long been a problem in asphalt mixtures and thus a great deal of research has been focused on the development of a rheological parameter that would address the rutting susceptibility of both unmodified and modified bituminous binders. In this research, three warm mix additives (Sasobit, Rheofalt and Zycotherm) were used to modify 60-70 penetration grade base binder. The rutting potential of both modified and unmodified binders were evaluated through the multiple stress creep recovery (MSCR)-based parameter, nonrecoverable compliance (J_nr) and recovery parameter (R). Several performance tests carried on stone matrix asphalt (SMA) mixtures comprising different nominal maximum aggregate sizes (NMASs, 9.5, 12.5 and 19 mm), like Marshall stability, dynamic and static creep and Hamburg wheel tracking tests to evaluate their rutting performance. The objective of this work is to correlate MSCR test results to performance. Results indicate that for the range of the gradations investigated in this work, increasing the nominal maximum aggregate size of the gradation would increase the permanent deformation resistance of the SMA mixture. Addition of 3% sasobit to base binder leads an increase in J_nr100 about 82%. Addition of 2% rheofalt to base binder leads an recovery increase of about 9.76 % and 27.44% in stress levels of 100 and 3200 Pa, respectively. The results reveal that rutting resistance of mixtures improves as J_nr decreases. The use of the MSCR test in the rutting characterization of bituminous binders is highly recommended based on the results of this work.

Key words: warm additives; multiple stress creep recovery; rutting; flow number; flow time; wheel track test

1 Introduction

Stone matrix asphalt (SMA) is a hot mix asphalt (HMA) consisting of a coarse aggregate skeleton and a high binder content mortar. SMA was developed in Germany during the mid-1960s and it has been used in Europe for more than 40 years to provide better rutting resistance and to resist studied tyre wear [1]. Because of the superior performance of SMA mixture comprising its high rut resistance, high skid resistance, high durability, improved resistance to reflective cracking, better drainage condition and reduced noise pollution [2], SMA has been widely adopted in Europe, Australia, USA, Canada, Japan and many other countries worldwide, as a surface course for heavily trafficked roads [3].

Rutting is one of the most important distresses for asphalt pavements. It is caused by material consolidation and lateral movement due to repeated heavy wheel loadings on the various pavement layers/subgrade. The distress is manifested by a depressed rut along the wheel path on the pavement surface. Different factors can influence rutting properties of asphalt mixture, including: aggregate type and gradation, amount of air void in asphalt mixture, type and amount of binder content, environmental temperature as well as mode and amount of loading applied on road pavement [4, 5]; the methods to prevent the rutting are primarily through engineering an asphalt mixture with improved shear resistance to withstand problems posed by the environment and traffic loadings [6].

The Superpave specification parameter, |G*|/sin δ, is identified as the term to be used for high temperature performance grading of paving asphalts in rating the binders for their rutting resistance. Although used for many years as a rutting parameter, it has been demonstrated that the relationship between |G*|/sin δ and rutting is poor. This term is found to be inadequate in describing the rutting performance of certain binders, particularly polymer modified binders. In NCHRP 9-10 project, BAHIA et al [7] evaluated the direct correlation between mixture’s rutting properties and |G*|/sin δ on RTFO aged binders, tested at the same temperature when the mixture test was conducted. The results indicated a poor correlation (R₂=23.77%) between the mixture rate of accumulated strain (S) and the parameter |G*|/sin δ measured at 10 rad/s [7]. Other researchers have confirmed this conclusion [8]. Consequently, many

agencies introduced additional tests to the standard PG specifications to overcome this limitation, known as PG-plus.

The multiple stress creep recovery (MSCR) is proposed as a better test to evaluate modified binders and estimate their role in pavement performance. [8].

In order to reduce the production temperature and consequently the energy consumption in the manufacture of asphalt mixtures, a new concept was developed and has been tested and implemented in the last few years. This is the warm mix asphalt (WMA) technology [9], which intends to lower the production temperature but keeps the mechanical and rheological properties of warm asphalt mixtures as close as possible to those of conventional mixtures [10, 11]. In order to increase their sustainability, it should be assessed by a life cycle analysis similar to those available in Refs. [12-14]. Moreover, the use of this technology leads to a decrease on the emissions of gases and odors from asphalt plants, and an improvement on the personnel working conditions [15-21].

The objective of this work is to correlate the MSCR parameter (J_nr) with resistance to permanent deformation based on flow number, flow time and wheel tracking tests. Also, the addition of warm mix additives into the asphaltic mixture can complicate the engineering process; more knowledge is needed to assess the influence of the warm additives to the pavement rutting performance. The research methodologies to achieve the objective are: perform binder testing to measure non-recoverable creep compliance (J_nr) and recovery (R), and superpave AASHTO M320 high temperature parameters; perform flow number, flow time and wheel tracking tests to measure the rutting resistance of asphalt mixes; correlate the binder results to flow number, flow time and wheel tracking tests results; make a case for or against the use of J_nr as valid test parameter for warm modified binders.

2 Materials and methods

2.1 Materials

Aggregates used in the present work were supplied from Asb-cheran Quarry located in eastern part of Tehran province. The physical and chemical properties of aggregates are presented in Tables 1 and 2, respectively. The particle size distribution of the aggregates is shown in Fig. 1 for SMA mixture types including types A, B and C with nominal maximum aggregate size (NMAS, S_nma) of 9.5, 12.5 and 19 mm, respectively.

Asphalt binder used in this work was AC-60/70, provided from Tehran Refinery and Pasargad Oil Company, Tehran, Iran. Physiochemical properties of the base binder are given in Table 3.

Cellulose fiber was added to the SMA mixtures as drain down inhibitors. The properties of the used fiber are provided in Table 4. As suggested by national cooperative highway research pavement (NCHRP) Report No. 425 [22], the added amount of cellulose fiber was 0.3% of the mixture (mass fraction). Fibers were uniformly mixed with aggregate before incorporating the asphalt cement.

Table 1 Engineering properties of aggregate

Table 2 Chemical composition of aggregate (mass fraction, %)

Fig. 1 Aggregate gradation chart:

Table 3 Properties of utilized bitumen considering related standards

Table 4 Properties of Cellulose fiber fibers

Three warm additive products were investigated in this research: sasobit, rheofalt LT70 and zycotherm. Sasobit is a long-chain aliphatic hydrocarbon obtained from coal gasification by the Fischer–Tropsch process. The Fischer–Tropsch process is a catalyzed chemical reaction in which carbon monoxide and hydrogen are converted into liquid hydrocarbons of various forms in the presence of iron and cobalt as catalysts. Sasobit forms a homo-geneous solution with the base binder upon stirring (1.5% of the binder, mass fraction), yielding marked reduction in the binder’s viscosity. After crystallization, sasobit forms a lattice structure in the binder, providing structural stability [23]. Sasobit is a wax product added at 1%-3% (mass fraction) of asphalt binder.

Rheofalt LT70, designed by Van Weezenbeek industries, is a mixture of paraffin waxes, synthetic resins of hydrocarbons, thermoplastic polymers and oxidation inhibitors, especially designed to enlarge the range of working temperature. This additive is designed to modify the bitumen properties and obtain a plasticizer effect to favor reduction of its viscosity, achieving manufacture temperatures between 20 °C and 30 °C lower than that in HMA, and laying temperatures around 70 °C, while maintaining the mechanical properties of HMA. This additive can be directly incorporated into the mixer or into the bitumen tank, requiring in this case a slight recirculation to achieve optimal dispersion. The amount to be added depends on the benefits required, although 2%-4% of bitumen (mass fraction) is recommended.

ZycoTherm is WMA additive developed by Zydex Industries, Gujarat, India. This is an odour free, chemical warm mix additive that has been engineered to provide significantly improved benefits over current WMA technologies by offering lower production and compaction temperatures, while simultaneously enhancing the moisture resistance of pavements by serving as an antistrip. Mixes that have been modified with ZycoTherm can be produced at 120-135 °C and compacted at 90-120 °C. Overall, ZycoTherm offers temperature reductions depending on the properties of the mix. ZycoTherm has built in antistrip mechanism that allows it to dually function as an antistrip as well as a warm mix additive. The additive is universally compatible with all types of modified as well as unmodified binders, including polymer modified bitumen and crumb rubber modified bitumen binders. It does not affect binder grading or change any other binder properties. Zyco therm is identified as chemical products at room temperature, and added at 0.1% and 0.15% (mass fraction) that of asphalt binder [24].

The physical and chemical properties of WMA additives are presented in Table 5. The mixing and compaction temperatures adopted for various mixture types are presented in Table 6.

The virgin and WMA-modified binders were tested and the test results are presented in Table 7. The binder without any additive is the control and referred as pure binder in this work.

The mineral filler used was limestone powder, which was passed through the #200 sieve. This large amount of filler (8%-10% of mixture, mass fraction) plays an important role in the properties of SMA mixture particularly in terms of air voids, voids in the mineral aggregate, and optimum asphalt content. Since the amount of material passing the 0.075 mm sieve is relatively large, the SMA mixtures perform very differently from other HMA mixtures.

2.2 Mix design

The mixture design method used for SMA was as

proposed by National Cooperative Highway Research Program (NCHRP) Report No. 425 [22]. The SMA design is based on the volumetric properties of the mixture. Some key criteria are air voids, voids in mineral aggregate and the VCA or voids in coarse aggregate ratio that is an indicator of the stone-on-stone contact. The later is the most important factor affecting the performance of SMA mixtures. In this wrok, three different SMA mixtures have been designed with different gradations. For each gradation, NMAS of 19 mm, 12.5 mm and 9.5 mm, three trial gradations were prepared. The three trial blends were along the coarse and fine limits of the gradation band along with one gradation falling in the middle. The next step was to check the aggregate skeleton by determination of voids in coarse aggregate in dry rodded condition. The test was carried out on the coarse aggregate fraction of any gradation. For instance, the coarse fraction is the portion of the total aggregate blend retained on the No. 4 sieve for the 12.5 mm and 19 mm NMAS SMA mixtures. For the 9.5 mm NMAS SMA mixture, the stone is the portion of aggregate blend retained on the No. 8 sieve. Two replicates of any gradation were tested and the average results were used. The test was carried out according to AASHTO T19. For each trial gradation, an initial trial asphalt binder content of 6.2% was selected and three specimens were compacted by 50 blows of Marshall hammer in accordance with ASTM D 1559 [26]. Seventy-five compaction blows were not used because this would not result in a significant increase in density over that provided by 50 blows. SMA mixtures have been more easily compacted to the desired density on the roadway than that with the effort required for conventional hot-mix asphalt mixtures [27]. The optimum gradations were then selected based on analysis results for Va, voids in mineral aggregate (VMA), void in coarse aggregate (VCA) ratio. The optimum gradation satisfies the minimum requirements in a way that provides VMA greater than 17%, VCA ratio less than or equal to 1, and Va between 3% and 4%. Based on the results, it is concluded that the optimum gradation falls in middle band for NMAS 19 mm, 12.5 mm and 9.5 mm, respectively.

Table 5 Physical and chemical properties of WMA additives

Table 6 Mixing and compaction temperatures adopted for WMA mixtures

Table 7 Binder test results

The next step in design of the SMA mixtures was to determine the optimum bitumen content. For this purpose, nine compacted replicates and three un- compacted ones were prepared for each gradation. The volumetric analyses were then conducted, and VCAmix, VMA and Va were determined. The following equations are mainly used for the purpose of volumetric analysis of the mixtures as

                          (1)

                      (2)

                               (3)

where G_mb is the bulk specific gravity of compacted specimen, G_ca is the combined bulk specific gravity of the coarse aggregate, γ_w is the density of water (998 kg/m³) V_drc is the voids in coarse aggregate in dry- rodded condition, V_mix is the VCA of the compacted mixture, P_bp is the fraction of aggregate by weight of the mixture remaining on the breaking point sieve, P_s is the fraction of aggregate in the mixture, and P_abp is the fraction of aggregate by the total weight of the aggregate remaining on the breaking point sieve [28].

Regarding the climatic regions in Iran that are suitable for construction of SMA courses, the air voids were designated to be 4% to avoid bleeding distress. From the analysis, the bitumen content associated with this Va was determined to be 6.9%, 6.8% and 6.5% for NMAS of 9.5 mm, 12.5 mm and 19 mm, respectively. This was used in preparing WMASMA mixtures to maintain consistency throughout the work. Three identical cylindrical (d10.16 cm×6.35 cm) samples for each mixture type and for each test (dynamic creep, static creep, and wheel tracking test) were fabricated.

2.3 Drain down

In order to evaluate the potential of drain down, a test procedure suggested by NCHRP Report No. 425 [22] was conducted for SMA mixtures. Thus, for each set of mixture, a standard wire basket containing uncompacted SMA sample along with a plate was placed in an oven at the mixing temperature. After about 1 h, the amount of binder drain down (A_bdd) was calculated by the following equation:

                            (4)

where A is the initial plate mass, B is mass of plate plus drained materials and W is the loose sample mass.

The binder drain-down results of different SMA mixtures are depicted in Fig. 2.

3 Experimental program

3.1 Rheological and physical bitumen tests

In this research, to study the properties of bitumens (unmodified and modified with WMA additives), rotational viscosity (RV) test and dynamic shear rheometer (DSR) test were conducted. Rotational viscosity test was conducted in order to be certain of bitumen pumping and bitumen mixing with hot aggregates. The results are shown in Table 1. The dynamic shear rheometer test at frequency of 10 rad/s (1.59 Hz) was conducted at required high temperature 64 °C and required intermediate temperature 25 °C to control permanent deformation at high temperatures and fatigue cracking at intermediate temperatures, respectively [29].

3.1.1 Multi-stress creep recovery (MSCR) test

Multi-stress creep recovery test, using the dynamic shear rheometer, was run on RTFO-aged samples. Anton Paar DSR with its parallel-plate geometry loading device and a control and data acquisition system were utilized for conducting the MSCR test in the present study. Specimens were tested in replicates using a 25 mm disc and with 1 mm gap setting at 64 °C and at stresses of 100 and 3200 Pa. The tests were performed at the selected temperatures using a constant stress creep of 1 s duration

and a relaxation period of 9 s, for ten cycles at each stress level. Recoverable and non-recoverable components of creep compliance were determined at the end of 10 cycles [30]. A typical schematic diagram of stress application and accumulation of strain in response to applied stress may be found elsewhere [31].

Fig. 2 Binder drain down of different SMA mixtures (P–pure bitumen; S–Sasobit; R–Rheofalt; Z–Zycotherm)

Initial strain (ε₀) value at the beginning of creep portion of each cycle and strain value at the end of creep portion (ε_c) of each cycle were determined. The difference in both strains is known as adjusted strain (ε₁). Similarly, strain values (ε_r) at the end of recovery portion of each cycle, and adjusted strain value (ε₁₀) at the end of recovery portion of each cycle, were computed. The adjusted strain value (ε₁₀), recovery, and non-recoverable compliance are calculated for ten cycles at each constant stress level using the following formulae [32]:

                                 (5)

                                 (6)

                                 (7)

where J_nr is the non-recoverable creep (1/kPa) as noted in Eq. (7), which measures the residual strain in a specimen after a creep and recovery cycle divided by the stress applied, and indicates the resistance of an asphalt binder to permanent deformation under repeated load [30]. In Eq. (7), γ_u,a is the unrecovered strain from the end of the 9 s recovery portion of the creep and recovery test, τ is the shear stress applied during the 1 s creep portion of the creep and recovery test. Creep and strain measurements were recorded for different stress levels. Creep recovery test determines the mechanical properties of asphalt binders and creep compliance of asphalt cement, which can be used to predict rutting of asphalt mixtures [31, 32].

The stress sensitivity of the asphalt binders was evaluated by means of the differences in non- recoverable compliances (J_nr,diff). This parameter shows the increase in the J_nr value of the asphalt binder when the stress level increases from 100 to 3200 Pa, as it can be seen in the equations in Fig. 3. In practical terms, it evaluates the susceptibility of the asphalt binder to rutting when unexpected heavy traffic loadings are applied on the pavement structure or unusually high temperatures are observed in the field [33, 34].

3.2 Performance tests

3.2.1 Marshall stability, flow and Marshall quotient

The Marshall properties of SMA mixtures were evaluated in this test according to ASTM D1559. The Marshall quotient (MQ, Q_m), which is an indicator of resistance against deformation of the bituminous mixture, is also calculated. MQ values can be used as a measure of the material’s resistance to shear stresses, permanent deformation and rutting in service. Higher MQ values indicate stiffer and more resistant mixtures [35].

3.2.2 Dynamic creep test

Rutting in asphalt concrete pavements under traffic loading occurs predominantly at elevated temperatures. Rutting in the asphalt surface layer is more sensitive to the resistance of the layer to shape distortion due to plastic flow than it is to the layer resistance to densification or volume change [36]. Based on the considerations, the dynamic creep test in this research was conducted in accordance with US. NCHRP 9-19 [37]. The test was conducted at 50 °C. The stress of 450 kPa was used as deviator stress with 0.1 s loading and 0.9 s rest time for each loading pulse cycle.

3.2.3 Static creep test

The test was performed as uniaxial static creep test, the specimen was subjected to a constant compressive load of 600 kPa at a test temperature of 52.5 °C (130 °F). The test may be conducted with or without confining

pressure. In this work, the test was performed without confining pressure.

Fig. 3 Plot of MSCR test and its final outcomes [31]

Flow time was conducted for 10000 s or until the sample failed due to cracking. The resulting axial strain was measured as a function of time and numerically differentiated to calculate the flow time which was defined as the time corresponding to the minimum rate of change of axial strain.

3.2.4 Wheel tracking test

Wheel-tracking devices are popular to estimate or evaluate the permanent deformation of asphalt pavement. The Hamburg wheel tracking device (HWTD), the georgia loaded wheel tester (GLWT), the asphalt pavement analyzer (APA), the LCPC (French) wheel tracker, the purdue university laboratory wheel tracking device (PUR Wheel), and the one-third scale model mobile load simulator (MMLS3) are the currently available devices in world-wide. Each wheel-tracking device has its own advantages and disadvantages. Advantages include similar rutting and deformation values of field performance for HMA. Disadvantages are different compactions and test devices lack of other pavement coarse type effects, steel wheel effects and single adopted axles [38].

The wheel-tracking test was conducted employing the Hamburg wheel-tracking device for evaluation of pavement performance in high temperature. Specimens mixed with the determined asphalt contents from mix design and fabricated by the rolling machine were of dimensions of 300 mm×300 mm in cross-sectional area and 50 mm in height at an air void ratio of about 7%, according to AASHTO-T324 [39]. The wheel tracking test was performed using 5.5×10⁵ Pa wheel pressure at 60 °C temperature under dry condition. The wheel shall make 22 passes across the specimen per minute. Rut depth of asphalt mixtures was measured for 20000 passes of 5.5 kg/cm² loaded wheels at 60 °C.

4 Results and discussions

4.1 Multiple stress creep and recovery (MSCR) test

The results of the MSCR test are presented in Table 8. Recoveries of the asphalt binders at two different stress level of 100 Pa and 3200 Pa are presented in Table 8. The addition of warm additives to a 60/70 base bitumen increases the recovery at common high pavement temperatures. Higher recoveries indicate that the asphalt binder can recover a higher portion of its total strain at the end of each loading–unloading cycle, which is favorable to the resistance of the material to rutting. As observed, the base binder has a few recoveries in stress levels of 100 and 3200 Pa. The Zycotherm-modified binder has the lowest recovery in any of the stress levels. It can be due to the low stiffness and the low elastic behavior of the sample. However, the 3% sasobit modified binder shows higher recovery in both stress levels which indicates that the sasobit modified binders are more rut resistant than other samples. On the other hand, addition of 2% rheofalt to base binder leads an increase of recovery of about 9.76 % and 27.44% in stress levels of 100 and 3200 Pa, respectively.

As Table 8 shows, the compliance (J_nr) in stress levels of both 100 and 3200 Pa, decreases as the sasobit content increases. So that, the binder containing 3% sasobit has the lowest J_nr value. Lower nonrecoverable compliances indicate a minor contribution of the asphalt binder to the appearance of rutting in the asphalt mixture or, in other words, a lower susceptibility of the bituminous material to rutting.

The results are much better for the sasobit-modified binders than for the rheofalt. As the results show, the zycotherm modified binder has the highest J_nr value. It can be concluded that zycotherm leads the pure binder more susceptible to rutting. Also, the J_nr value first increases by addition of 2% rheofalt and then decreases by addition of 4% rheofalt.

As previously observed for the recovery, lower non-recoverable compliances also suggest that the modified asphalt binders are less prone to rutting after the application of loading–unloading cycles at high pavement temperatures. In terms of the asphalt mixture, asphalt binders with lower J_nr and/or higher recovery less contribute to the accumulation of unrecovered strain in the asphalt layer. From the point of view of rheology, these lower J_nr values may be obtained by decreasing the

amount of unrecoverable strain at the end of the creep- recovery cycle, for a particular stress level. Since the stress levels are the same for all MSCR tests, it can be concluded that the amount of permanent strain is lower for the modified asphalt binders than for the unmodified one.

Table 8 Average values of MSCR test parameters calculated from collected test data

To calculate the amount of the sensitivity of the binder behavior to the imposed stress level, a parameter which was not measurable in previous methods, can be evaluated as

                      (8)

where J_nr,100 and J_nr,3200 are non-recoverable compliances under 100 and 3200 Pa, respectively.

The differences in non-recoverable compliances (J_nr,diff) at the creep and recovery test are shown in Table 8. This difference is a measure of the sensitivity of the asphalt binder to the increase in the stress level; therefore, lower values are associated with a less stress sensitive material. A revised version of the Superpave specification [40] establishes a maximum value of 75% for the parameter J_nr,diff at the high PG grade and the loading–unloading times of 1-9 s. Such criterion is able to eliminate asphalt binders that are overly sensitive to stress and potentially susceptible to rutting in the field, even though the other criteria of the performance-grade specification are met.

It can be seen from Table 8 that the value of the J_nr,diff parameter is particularly high for 3% Sasobit modified binder. Therefore, this binder can be considered to be higher stress sensitive than other additives at the creep stress of 3200 Pa. The pronounced stress sensitivity of binder containing Sasobit is further highlighted in Figs. 4 and 5 where the development of accumulated strain is presented for the selected modified binders separately during the first and last ten creep- recovery cycles of the MSCR test (where creep stresses of 100 and 3200 Pa are used, respectively) [20].

Fig. 4 Development of accumulated strain or selected modified binders during first ten creep-recovery cycles of MSCR test conducted at 100 Pa creep stress level

Fig. 5 Development of accumulated strain or selected modified binders during last ten cycle creep-recovery cycles of MSCR test conducted at 3200 Pa creep stress level

The accumulated strain at the end of the MSCR test (γ_acc) is depicted in Table 8. It is observed that addition of 1% sasobit, 3% sasobit and 4% rheofalt decreases the accumulated strain by 28.65%, 64.86% and 13.42%, respectively. On the other hand, addition of 2% rheofalt and zycotherm leads an increase by 6.3 % and 17.12 %, respectively. The values of J_nr 3200 and γ_acc have a very strong linear correlation as depicted in Fig. 6. This is due to the fact that the accumulated strain at the end of the test is created almost completely during the last ten creep-recovery cycles of the test. Because of the applied creep stress, the resulting creep deformation is much higher during the latter half of the test than that during the first half [20].

Fig 6 Correlation between J_nr3200 and γ_acc parameter values

4.2 Correlation identification between oscillation- based rutting parameter of G*/sinδ and MSCR parameter

The SHRP uses oscillation-based rutting parameter of G*/sinδ to measure rutting resistance of asphalt binders. Meanwhile, based on new loading approach of creep recovery and the calculation of nonrecoverable compliance J_nr, the MSCR test entirely can obviously capture the binder nonrecoverable response, which is always related to the permanent deformation. Figure 7 shows the relationships between MSCR-based parameter and the oscillation-based parameter (G*/sin δ), which is inversed to have the same unit 1/kPa as J_nr [41].

Fig. 7 Relationship between Oscillation-based G*/sinδ and MSCR-based J_nr

4.3 Dynamic creep test results

Total permanent strains of control and warm modified SMA mixtures were calculated and compared with each other, the results are illustrated in Fig. 8. Aggregate interlock and bitumen stability are two important factors affecting the creep performance of asphalt concrete. Modification of binder with Sasobit stiffens the base binder. As can be seen, unmodified mixture explicitly indicates the less resistance to permanent deformation; however, Sasobit-modified mixtures could be the most rut resistant mixture among all SMA samples, and in all cases, total permanent strain decreases at higher Sasobit content (3%) in three NMAS. The increase in the strength of mixtures by increasing the amount of Sasobit can be justified by the crystallization phenomenon of wax in bitumen at temperatures lower than melting point of wax. This process influences the results of dynamic creep behavior. As the binder test results show, binders containing sasobit has the lowest non-recoverable compliance (J_nr) and highest recovery value. The lower the J_nr and higher the recovery, the better the resistance of binder to rutting. This confirms the results of rutting test on mixtures. Mixture containing 3% sasobit with NMAS of 19 mm has better resistance to permanent deformation than mixtures with NMAS of 9.5 and 12.5 mm containing other warm additives. Increasing the NMAS of mixtures containing 1% Sasobit from 9.5 to 19 mm leads to the increase in flow number of mixtures by 20% and 32%, respectively. From Fig. 8, it could be observed that different contents of Rheofalt are relatively capable of increasing the resistance of SMA mixtures to rutting. The role of 4% Rheofalt was more significant than 2% content. In NMAS of 9.5 and 19 mm, increasing the Rheofalt content from 2% to 4% increases the flow number of mixtures about 10%. But in NMAS of 12.5 mm, increasing the Rheofalt content from 2% to 4% increases the flow number of mixtures about 30%. The binders modified with 4% Rheofalt have lower accumulated strain (γ_acc) than binders modified with 2% Rheofalt. As the binder results shown in Table 8, the Zycotherm-modified binder has the highest γ_acc between modified binders. Also Zycotherm-modified binders have a higher J_nr and lower recovery than others. This confirms the results of dynamic creep test in Fig. 8, in which the mixtures containing Zycotherm have the lowest flow number. For the range of the gradations investigated in this works, it could be recognized that increasing the nominal maximum aggregate size of the gradation would increase the permanent deformation

resistance of the SMA mixture.

Fig. 8 Flow number results of Asphalt mixtures

As the binder tests results show, the binders containing warm additives have better recovery than base binder. Higher recoveries indicate that the asphalt binder can recover a higher portion of its total strain at the end of each loading–unloading cycle, which is favorable to the resistance of the material to rutting. As the results of dynamic creep test show, mixtures containing warm additives have better resistance to deformation.

4.4 Static creep test results

As shown in Fig. 9, mixtures containing 3% Sasobit with NMAS of 12.5 and 19 mm register the highest flow time value as compared to the remaining mixtures. As results show, mixtures containing Sasobit have the highest flow time among each NMAS. MSCR test results also show that the binders containing Sasobit have the lowest J_nr, which indicates the higher resistance of mixtures modified with Sasobit. The unmodified mixtures and mixtures with Zycotherm reach the lowest flow time value, the MSCR test results also indicate that their J_nr value are close together and are the highest. From Fig. 9, it could be observed that the Rheofalt additive has significant effect on increasing the resisitance of mixtures to rutting. The mixtures containing 4% Rheofalt have a better performance than 2% Rheofalt modified mixtures. The binders modified with 4% Rheofalt have lower accumulated strain (γ_acc) and J_nr and higher recovery (R) than binders modified with 2% Rheofalt. Addition of Rheofalt content from 2% to 4% leads an increase in rutting resistance of mixtures about approximately 10%. In static creep test, addition of Zycotherm to the mixture does not have any effect on increasing rutting resistance of mixtures.

The results indicate that by increasing the NMAS of mixtures, the flow time increases. In major NMAS mixtures, the interlock between aggregates is better, so the rutting resistance of mixture is enhanced. SMA is specially designed to rely on aggregate-to-aggregate interlock and since aggregate is much stronger, it is sensible that the mix would be as well.

4.5 Results of wheel tracking test and analysis

Data obtained from wheel tracker machine, as shown in Fig. 10, reveal that the asphalt mix with 3% Sasobit offers maximum rut resistance, and mixtures with Zycotherm and no additive show minimum rut resistance in any NMAS’s. As the binder test results show, binders containing Sasobit has the lowest non-recoverable compliance (J_nr) and highest recovery value. The lower the J_nr and the higher the recovery, the better the resistance of binder to rutting. This confirms the results of rutting test on mixtures.

Addition of Sasobit to base binder more stiffens the base binder in comparison with other warm additives. From Fig. 10, it could be observed that different contents of Rheofalt are relatively capable of increasing the resistance of SMA mixtures to rutting. The role of 4% Rheofalt is more significant than content of 2%. For the range of the gradations investigated in this work, it could be recognized that increasing the nominal maximum aggregate size of the gradation would increase the permanent deformation resistance of the SMA mixture. The reason for this is that stronger aggregate skeleton is formed in higher NMAS mixtures, increasing the internal friction and the load-bearing capacity among the

aggregates of these mixtures and hence minimize the potential for rutting.

Fig. 9 Estimated flow time of mixtures

Fig. 10 Rut depth of asphalt mixtures

4.6 Correlation between non-recoverable creep compliance (J_nr) and asphalt mixture rutting parameters

The main aim of this part is to suggest correlation between the non-recoverable compliance (J_nr) and mixtures’ rutting parameters value.

A comparison between rutting performance was evaluated with an accelerated loading facility (ALF) and the J_nr had an R² value of over 0.80—a strong correlation between the mix performance in the field and the binder performance in laboratory tests [42-46].

This shows that J_nr is a good indicator of rut prediction of asphalt mixtures. The results of flow number and flow time of different mixtures are compared with J_nr (at 3.2 kPa) values. As shown in Figs. 11 and 12, the final J_nr (at 3.2 kPa on x-axis) flow number and flow time (y-axis) of asphalt mixtures prepared with warm modified binders are depicted. As seen in Figs. 11 and 12, flow number and flow time of mixtures decrease as J_nr increases. Figures 11 and 12 also show a strong relationship between non-recoverable creep compliance of asphalt binder and flow number (R²=0.86) and flow time (R²=0.89) of modified asphalt mixtures.

The final rut depth, obtained from the wheel tracker, was compared with J_nr (at 3.2 kPa) values. Figure 13 shows a typical plot of final J_nr (at 3.2 kPa on x-axis) rut depth (y-axis) of asphalt mixtures prepared with warm modified binders. Figure 13 also shows a strong relationship (R²=0.89) between non-recoverable creep compliance of asphalt binder and rut depth of asphalt mix obtained from a wheel tracker test.

The results reveal that non-recoverable creep compliance of Asphalt binder can be used to predict therut resistance of Asphalt mixtures. Based on this index, asphalt mixtures can be designed using suitable grade of asphalt binder for different traffic levels and environmental conditions. It will be a useful index to predict the rut resistance of asphalt mixtures in the field.

Fig. 11 J_nr test results plotted versus flow Number results

Fig. 12 Correlation between J_nr test results versus flow time results

Fig. 13 Correlation between J_nr test results versus rut depth results

4.7 Data analysis method

A two-factor analysis of variance (ANOVA) was performed to check the significance of the individual factors on the design objectives. The rutting parameter was considered as the dependent variable. The two fixed factors considered are additive content and NMAS. The hypothesis is that there is no significant effect of the two factors on the measured engineering properties. The significance is tested considering 95% confidence interval. The dependent variables are significantly influenced by additive content and NMAS and the interaction between them.

Statistical analysis shows that the additives and NMAS have a significant effect on the mixtures rutting performance.

Table 9 Analysis of variance for Rheofalt additive in flow number test

Table 10 Analysis of variance for Rheofalt additive in flow time test

Table 11 Analysis of variance for Rheofalt additive in rut depth test

Table 12 Analysis of variance for Sasobit additive in flow number test

Table 13 Analysis of variance for Sasobit additive in flow time test

Table 14 Analysis of variance for Sasobit additive in rut depth test

5 Conclusions

1) According to MSCR test results, the addition of warm additives to a 60/70 original bitumen increases the recovery at common high pavement temperatures. The Zycotherm-modified binder has the lowest recovery in any of the stress levels. 3 % sasobit modified binder shows more recovery in both stress levels which indicates that the sasobit modified binders are more rut resistant than other samples.

2) The Zycotherm modified binder has the highest J_nr value. The results of J_nr value are much better for the sasobit-modified binders than for the Rheofalt. Although the recoveries are lower and the non-recoverable compliances are higher for the 3% Sasobit at any stress levels, the stress sensitivity is much higher for this additive than for the other warm additives.

3) For the range of the gradations investigated in this work, it could be recognized that increasing the nominal maximum aggregate size of the gradation would increase the permanent deformation resistance of the SMA mixture. Mixtures containing 3% Sasobit have better rutting resistance than other mixtures through dynamic creep, static creep and wheel track tests.

4) Non-recoverable creep compliance captures rut resistance of asphalt mixtures. Reasonable relationship exists (R²=0.89) between the rut depth of asphalt mixtures obtained from wheel tracker machine, and J_nr of an asphalt binder.

5) The non-recoverable compliance J_nr has been demonstrated to correlate with rutting performance, as measured by flow number and flow time testing, with rutting resistance improving as J_nr decreases. The use of the MSCR test in the rutting characterization of bituminous binders is highly recommended based on the results of this research.

References

[1] RIMSA V, KACIANAUSKAS R, SIVILEVICIUS H. Finite element simulation of the normal interaction of particles in the visco-elastic solid [J]. Mechanics and Control, 2011(30): 245-253.

[2] MOGHADAS N F, AFLAKI E, MOHAMMADI M A. Fatigue behavior of SMA and HMA mixtures [J]. Constr Build Mater, 2010, 24(7): 1158-65.

[3] SENGUL C E, ORUC S, ISKENDER E, AKSOY A. Evaluation of SBS modified stone mastic asphalt pavement performance [J]. Construction and Building Materials. 2013, 30, 41: 777-83.

[4] DELGADILLO R, BAHIA H U. The relationship between nonlinearity of asphalt binders and asphalt mixture permanent deformation [J]. Road Materials and Pavement Design, 2010, 11(3): 653-80.

[5] CHRISTENSEN JR D W, BONAQUIST R. Rut resistance and volumetric composition of asphalt concrete mixtures (With Discussion) [J]. Journal of the Association of Asphalt Paving Technologists, 2005: 74.

[6] KENNEDY T W, HUBER G A, HARRIGAN E T, COMINSKY R J, HUGHES C S, Von QUINTUS H, MOULTHROP J S. Superior performing asphalt pavements (Superpave): The product of the SHRP asphalt research program [R]. Washington D C, USA: National Research Council, 1994.

[7] BAHIA H U, HANSON D I, ZENG M, ZHAI H, KHATRI M A, ANDERSON R M. Characterization of modified asphalt binders in superpave mix design [M]. Washington D C, USA: Transportation Research Board, 2001.

[8] GOLALIPOUR A. Modification of multiple stress creep and recovery test procedure and usage in specification [D]. Madison, USA: University of Wisconsin–Madison, 2011.

[9] RUBIO M C, G, BAENA L, MORENO F. Warm mix asphalt: An overview [J]. Journal of Cleaner Production, 2012, 24: 76-84.

[10] ZHAO W, XIAO F, AMIRKHANIAN S N, PUTMAN B J. Characterization of rutting performance of warm additive modified asphalt mixtures [J]. Construction and Building Materials, 2012; 31: 265-272.

[11] SILVA H M, OLIVEIRA J R, PERALTA J, ZOOROB S E. Optimization of warm mix asphalts using different blends of binders and synthetic paraffin wax contents [J]. Construction and Building Materials, 2010, 24(9): 1621-1631.

[12] HUANG Y, BIRD R, HEIDRICH O. Development of a life cycle assessment tool for construction and maintenance of asphalt pavements [J]. Journal of Cleaner Production, 2009, 17(2): 283-296.

[13] CHIU C T, HSU T H, YANG W F. Life cycle assessment on using recycled materials for rehabilitating asphalt pavements [J]. Resources, Conservation and Recycling. 2008, 52(3): 545-556.

[14] WHITE P, GOLDEN J S, BILIGIRI K P, KALOUSH K. Modeling climate change impacts of pavement production and construction [J]. Resources, Conservation and Recycling. 2010, 54(11): 776-82.

[15] HURLEY G C, PROWELL B D, HUNER M. Evaluation of Aspha-min Zeolite for use in warm mix asphalt [R]. Alabama, USA: National InCenter for Asphalt Technology, Auburn University, 2005.

[16] HURLEY G C, PROWELL B D. Evaluation of sasobit for use in warm mix asphalt [R]. Auburn, USA: Auburn University, 2005.

[17] PROWELL B D. Warm Mix Asphalt [R]. Alexandria, USA: American Trade Initiatives, 2007.

[18] PROWELL B. Warm mix asphalt [R]. Alexandria, USA: American Trade Initiatives, 2007.

[19] OLIVEIRA J R, SILVA H M, ABREU L P, FERNANDES S R. Use of a warm mix asphalt additive to reduce the production temperatures and to improve the performance of asphalt rubber mixtures [J]. Journal of Cleaner Production, 2013, 41: 15-22.

[20] LAUKKANEN O V, SOENEN H, PELLINEN T, HEYRMAN S, LEMOINE G. Creep-recovery behavior of bituminous binders and its relation to asphalt mixture rutting [J]. Materials and Structures, 2015, 48(12): 4039-4053.

[21] DUBOIS E, MEHTA Y, NOLAN A. Correlation between multiple stress creep recovery (MSCR) results and polymer modification of binder [J]. Construction and Building Materials, 2014, 65: 184-190.

[22] BROWN E R, COOLEY L A. Designing stone matrix asphalt mixtures for rut-resistant pavements [R]. Washington, USA: Transportation Research Board, 1999.

[23] SASOBIT TECHNOLOGY. Sasol wax [EB/OL]. [2011-01-18]. http://www.sasolwax.com/ Sasobit_Technology.html.

[24] ROHITH N, RANJITHA J. A study on marshall stability properties of warm mix asphalt using zycotherm a chemical additive [J]. International Journal of Engineering Research and Technology, 2013, 2(7): 19-27.

[25] ASTM D92. Standard test method for flash and fire points by cleveland open cup tester [S]. 2012.

[26] ASTM D1559. Standard test method for resistance of plastic flow of bituminous mixtures using Marshall apparatus [S]. 2003.

[27] ASTM D7405-10a. Standard test method for multiple stress creep and recovery (MSCR) of asphalt binder using a dynamic shear rheometer [S]. 2010.

[28] FAKHRI M, KHEIRY P T, MIRGHASEMI A A. Modeling of the permanent deformation characteristics of SMA mixtures using discrete element method [J]. Road Materials and Pavement Design. 2012, 13(1): 67-84.

[29] AMERI M, MANSOURIAN A, SHEIKHMOTEVALI A H. Laboratory evaluation of ethylene vinyl acetate modified bitumens and mixtures based upon performance related parameters [J]. Construction and Building Materials, 2013, 40: 438-447.

[30] AASHTO TP70. Standard practice for multiple stress creep recovery test of asphalt binder using a dynamic shear rheometer [S]. 2010.

[31] WASAGE T L, STASTNA J, ZANZOTTO L. Rheological analysis of multi-stress creep recovery (MSCR) test [J]. International Journal of Pavement Engineering, 2011, 12(6): 561-568.

[32] HAFEEZ I, KAMAL M A. Creep compliance: A parameter to predict rut performance of asphalt binders and mixtures [J]. Arabian Journal for Science and Engineering, 2014, 39(8): 5971-5978.

[33] ANDERSON M, D'ANGELO J, WALKER D. MSCR: A better tool for characterizing high temperature performance properties [J]. Asphalt, 2010, 25(2): 34-46.

[34] D'ANGELO J, KLUTTZ R, DONGRE R N, STEPHENS K, ZANZOTTO L. Revision of the superpave high temperature binder specification: The multiple stress creep recovery test (with discussion) [J]. Journal of the Association of Asphalt Paving Technologists, 2007, 76(8): 123-162.

[35] ZOOROB S E, SUPARMA L B. Laboratory design and investigation of the properties of continuously graded Asphaltic concrete containing recycled plastics aggregate replacement (Plastiphalt) [J]. Cement and Concrete Composites, 2000, 22(4): 233-242.

[36] BAHUGUNA S. Permanent deformation and rate effects in asphalt concrete: constitutive modeling and numerical implementation [D]. Cleveland, USA: Case Western Reserve University, 2003

[37] WITCZAK M W. Specification criteria for simple performance tests for rutting [J]. Transportation Research Board, 2007, 1: 63-98.

[38] LEE K H, CHO Y H. Design and performance of the HEART wheel load simulator [J]. Journal of testing and evaluation, 2003, 31(6): 1-7.

[39] AASHTO T324. Standard method of test for hamburg wheel track testing of compacted hot mix asphalt (HMA) [S]. 2011.

[40] AASHTO M320-09. Standard specification for performance-graded asphalt binder [S]. 2009.

[41] WANG C, ZHANG J. Evaluation of rutting parameters of asphalt binder based on rheological test [J]. International Journal of Engineering and Technology, 2014, 6(1): 30.

[42] ZOOROB S E, SUPARMA L B. Laboratory design and investigation of the properties of continuously graded Asphaltic concrete containing recycled plastics aggregate replacement (Plastiphalt) [J]. Cement and Concrete Composites. 2000, 22(4): 233-242.

[43] L Song-tao, ZHENG Jian-long. Normalization method for asphalt mixture fatigue equation under different loading frequencies [J]. Journal of Central South University, 2015, 22(7): 2761-2767.

[44] HASAN Z, HAMID B, AMIR I, DANIAL N. Long term performance of warm mix asphalt versus hot mix asphalt [J]. Journal of Central South University. 2013, 20(1): 256-266.

[45] LIAO Gong-yun, YANG Yi-wen, HUANG Xiao-ming, XIANG Jin-yuan. Permanent deformation response parameters of asphalt mixtures for a new mix-confined repeated load test [J]. Journal of Central South University, 2013, 20(5): 1434-1442.

[46] MA Tao, WANG Zhen, ZHAO Yong-li. Degradation behavior of aggregate skeleton in stone matrix asphalt mixture [J]. Journal of Central South University of Technology, 2011, 18(6): 2192-2200.

(Edited by FANG Jing-hua)

Cite this article as: Rezvan Babagoli, Hassan Ziari. Evaluation of rutting performance of stone matrix asphalt mixtures containing warm mix additives [J]. Journal of Central South University, 2017, 24(2): 360-373. DOI: 10.1007/s11171- 017-3438-z.

Received date: 2015-11-24; Accepted date: 2016-02-16

Corresponding author: Rezvan Babagoli, PhD candidate; Tel/Fax: +98-2177240565; E-mail: Rezvan_babagoli@civileng.iust.ac.ir