- Email: [email protected]
Mechanistic-empirical pavement performance of asphalt mixtures with recycled asphalt shingles
Construction and Building Materials 160 (2018) 687–697
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Mechanistic-empirical pavement performance of asphalt mixtures with recycled asphalt shingles Sharareh Shirzad a, Max A. Aguirre a, Luis Bonilla a, Mostafa A. Elseifi b,⇑, Samuel Cooper c, Louay N. Mohammad b a b c
Louisiana State University, United States Department of Civil and Environmental Engineering, Louisiana State University, United States Louisiana Transportation Research Center, United States
h i g h l i g h t s The effect of recycled asphalt shingles on pavement performance was evaluated. Pavement Mechanistic-Empirical (ME) predicted rutting performance correctly. There was an inconsistency in fatigue cracking prediction for mixes with RAS.
a r t i c l e
i n f o
Article history: Received 26 September 2017 Received in revised form 17 November 2017 Accepted 20 November 2017 Available online 24 November 2017 Keywords: Asphalt mixture Recycled asphalt shingle Mechanistic-empirical pavement design Recycling agents Semi-circular bending
a b s t r a c t As highway agencies are in the process of adopting the new mechanistic-empirical pavement design guide, Pavement ME, it is unclear how asphalt mixtures incorporating Recycled Asphalt Shingle (RAS) will influence the design when mechanistic-empirical approaches are used. In this study, Pavement ME was used to evaluate the effects of RAS (with or without recycling agents [RAs]) on pavement performance. Furthermore, a cost analysis was conducted to assess the life-cycle cost of asphalt pavements constructed with RAS. Three different pavement structures were analyzed at three traffic levels (low, medium, and high) and for two climatic regions (cold and hot). Pavement ME predicted that the mix with 5% PostConsumer Waste Shingle (PCWS) and 5% RA would be the best performer against roughness, rutting, and fatigue cracking. This is due to the stiffening effect of RAS, which is reflected in the dynamic modulus inputted in the software. While one would expect RAS to improve rutting performance, the superior fatigue performance of mixes with RAS was not expected given that the binder in RAS is an air-blown asphalt binder with poor elongation and relaxation characteristics. Results were compared to the Semi-Circular Bending (SCB) test results, which realistically predicted that the mixes with RAS and recycling agents would be the worst performers against cracking. This can be explained by the increased availability of aged RAS binder in these mixes when a recycling agent is used. Results of the cost analysis showed that the mixtures with RAS are more economical to produce. When considering the predicted performance of the mixes, the mix with 5% PCWS and 5% RA had the lowest cost over the pavement service life. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction The sustainability movement in road applications aims to achieve production, distribution, and construction of asphalt mixtures designed to last longer with less impact on ecological systems. Sustainable pavements would also minimize the use of natural resources, reduce energy consumption and greenhouse gas emissions, limit pollution, improve health and safety, and ⇑ Corresponding author. E-mail address: [email protected] (M.A. Elseifi). https://doi.org/10.1016/j.conbuildmat.2017.11.114 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
ensure a high level of user comfort [1]. Utilizing recycled materials to decrease the amount of virgin asphalt binder or aggregate used in the production of asphalt mixtures is an example of sustainable pavement technologies. Studies have shown that using recycled materials, such as recycled asphalt shingles (RAS), reduces the environmental impacts related to extraction, transportation, and processing of virgin materials [2]. Different recycling agents (RAs) may also be used in the asphalt mixtures containing RAS to improve the blending between recycled and virgin materials [3]. Recycling agents may also have the ability to reverse the aging process by softening or
688
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
rejuvenating the aged binder [4]. In addition to the environmental benefits, cost reduction associated with the use of RAS as well as the cost increase due to the use of recycling agents are important factors that should be evaluated when studying asphalt mixtures incorporating RAS and RAs. Traditionally, pavements were designed based on empirical principles, which used road tests results conducted by the American Association of State Highway Officials (AASHO) in the late 1950s and early 1960s in Ottawa, Illinois. The limitations of empirical design include that: (1) the design considers pavement performance through a subjective rating; (2) it offers limited attention to failure modes; and (3) it depends on conditions remaining the same, similar to the original field conditions. Yet, a 54% increase in the number of vehicles using the roads and a 75% increase in the vehicle miles traveled was observed between 1973 and 1993; thereby making today’s traffic loads significantly higher than the ones utilized to develop the semi-empirical 1993 AASHTO design method [5]. Based on these limitations, a Mechanistic-Empirical (ME) design method was developed, which accounts for variability in material properties, traffic loads, and construction procedures. In this new approach, both mechanical and empirical methods are incorporated by calculating pavement responses due to loading then relating those responses to pavement performance. Previous studies have evaluated the mechanistic properties, economic benefits, and ecological impacts from sustainable pavement technologies [6–15]. Yet, the incorporation of RAS and RAs in pavement design has not been thoroughly evaluated using ME design approaches. 2. Objectives and scope This study had two main objectives: (1) to evaluate the effects of RAS (with and without recycling agents) on the predicted performance from Pavement ME; and (2) to assess the life cycle cost of pavement constructed with RAS and recycling agents. An evaluation of the effects of RAS and RAs was conducted using Pavement ME for two different climatic conditions (cold and hot regions). Furthermore, the analysis encompassed three traffic levels (low, medium, and high) to evaluate the effect of traffic load on the performance of RAS materials. Level 1 inputs were used to describe the asphalt layers while Level 2 inputs were used to describe the granular base and subgrade layers. 3. Background New methodology, developed as part of NCHRP Project 1-37A, (Development of the 2002 AASHTO Guide for Design of New and Rehabilitated Pavement Structures: Phase II), sought to relate pavement distresses to pavement responses (stress, strain and deflection); this process is based on traffic loading, climate data, and material properties. These distresses may be used to predict pavement damage over time. Pavement ME is based on an iterative process, which initiates the analysis with a trial design and then evaluates the distress susceptibility of the primary design. If the predicted distresses meet the specified performance criteria, the primary design is accepted; otherwise, the evaluation process is repeated with a revised design. To utilize available data in the most efficient way, three different levels of input are considered in Pavement ME. This approach provides the user with the ability to set the level of input data according to pavement importance. Level 1 inputs evaluate pavement performance with respect to the mechanistic properties of the different layers, which in turn can be measured through laboratory or field testing. Therefore, Level 1 input has the highest accuracy, coupled with the highest expenses. Level 2 input addresses those predicted mechanistic val-
ues, which may be obtained from different databases or extrapolated from less extensive tests. Level 3 input considers the typical default values and therefore displays the lowest accuracy. Interestingly, the computational methodology used for distress prediction is the same for all input levels [16]. 3.1. Recycled asphalt shingles Recycled asphalt shingles, which represent one of the largest municipal solid wastes, can be used as a virgin binder replacement in asphalt pavements [17]. Every year, 11 million tons of waste shingles are produced, which results in 22 million cubic yards of waste materials, which must be landfilled [18]. Nineteen to 36% of RAS consist of a relatively hard asphalt binder; 2–15% fiberglass or cellulose backing, 20–38% fine aggregate, ceramic-coated natural rock, and the remainder of 1–12% is a mineral filler or stabilizer [19]. Reusing this waste material in asphalt pavement construction could present significant environmental and economic benefits. Starting from the 1980s, researchers evaluated the use of RAS in HMA production. Several studies were conducted to evaluate the impact of RAS on asphalt mixture performance [20–25]. Anurag et al. utilized the indirect tensile strength test to determine the effect of roofing waste with polyester fibers with different fiber lengths and different fiber contents on the moisture sensitivity of the mix. The study concluded that polyester fibers could improve the wet tensile strength and tensile strength ratio of the mixture [26]. More recently, Maupin used 5% RAS in asphalt mixture to evaluate the effect of RAS on the mix durability [27]. Results from the extracted binder material showed that the high-temperature grading of the binder was improved, while the low-temperature grading was not affected. Maupin also used mixture laboratory testing to evaluate the rutting, cracking, and tensile strength of the asphalt mixtures. Results of rut testing indicated that the mixes would perform satisfactorily on high traffic conditions. Similarly, the mixes were expected to perform satisfactorily against fatigue failure. Cooper examined the performance of different mixtures containing RAS with and without recycling agents [28]. The tests encompassed laboratory mechanistic tests to characterize the low, intermediate, and high temperature performance of mix, including the dynamic modulus test, semi-circular bending (SCB) test for intermediate temperature, thermal stress restrained specimen tensile strength test (TSRST) for low temperature, and a Hamburg type Loaded Wheel Tracking (LWT) test to evaluate rutting resistance. Results showed that although RAS caused a reduction in virgin binder, the extracted binder did not blend completely with the virgin binder. While there was no improvement in intermediate temperature cracking, the rutting resistance was improved [28]. Robinette and Epps studied the influence of RAS on energy consumption, emissions generation, and natural resource consumption [29]. They also reviewed the price of construction for mixtures containing RAS as compared to conventional materials and construction. Results indicated that in most cases, RAS addition could reduce energy consumption, generated emissions, and preserve natural resources such as aggregate and asphalt binder. By virtue of these processes, RAS can reduce the price of asphalt mixture production and construction as well as alleviating the overall environmental impacts. Arnold et al. examined the low-temperature cracking characteristics of asphalt mixtures containing various amount of RAS (0– 12.5%) using the Disc-shaped Compact Tension (DCT) and acoustic emission (AE) tests. Result showed that the mixtures with RAS had lower fracture energies and higher peak loads. The increase in peak load with the increase in RAS content indicated that the use of RAS did not affect the HMA’s high-temperature performance. However,
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
it significantly affected the embrittlement temperature of the mixtures [30]. Cascione et al. tested permanent deformation, fatigue cracking, and low temperature cracking performance of asphalt mixtures with RAS from seven different transportation agencies. When RAS was added to the mixtures, the high and low temperature grades of the virgin binder were increased. Furthermore, based on the flow number and dynamic modulus results, it was concluded that RAS addition improved the rutting resistance of the mixtures. Performance of all the mixtures, with or without RAS, was acceptable with respect to fatigue cracking. Lowtemperature fracture energy was measured for SCB samples and results suggested that RAS could increase the fracture resistance of mixtures prior to long term aging [31]. In another attempt to evaluate the effect of recycled materials on the performance of asphalt mixtures, Ghabchi et al. [32] examined the fatigue and low-temperature cracking potential of mixtures with RAS and/or RAP. Results indicated that the addition of 5% RAS and 5% RAP to the mixtures with a straight asphalt binder led to a maximum increase in the fatigue life of the mixtures. However, increasing the amount of RAS beyond 5% caused a decrease in the fatigue life of the mixtures. The results obtained from the creep compliance test indicated that RAS addition increased lowtemperature cracking potential [32].
689
naphthenic oil (referred to as RA1) and a vegetable oil derived from the pyrolysis of pine tree (referred to as RA2). As shown in Table 1, the first mixture was the conventional mixture (70CO) containing no RAS or RAs. The second mixture (70PG5P) contained 5% PostConsumer Waste Shingle (PCWS) and no RAs. Similarly, the third mixture (70PG5M) contained 5% Manufacturer Waste Shingle (MWS) and no RAs. The fourth mixture (70PG5P5HG) contained 5% PCWS with 5% RA2. The fifth mixture (70PG5P12CYCL) contained 5% PCWS with 12% RA1. Optimum dosage of the recycling agents was selected by utilizing volumetric and densification criteria. The aggregate gradation and virgin asphalt cement percentage were kept approximately constant while the percentage of the recycling agent was increased. This was necessary for the comparison of the control mixture with mixtures containing recycling agents. Furthermore, the asphalt contribution by the recycled materials was estimated after identifying the optimum virgin asphalt content. Table 2 presents gradation and job mix formula of the mixtures evaluated in this study. It also demonstrates the amount of asphalt binder contribution from RAS and the recycled binder ratio (RBR) in mixtures 70PF5P, 70PG5M, 70PG5P5HG and 70PG5P12CYCL. More details on the selection of the RAs’ contents have been presented elsewhere [28]. 4.3. Experimental test matrix
4. Methodology 4.1. Pavement performance prediction In this study, Pavement ME was used to predict the performance of asphalt mixtures with and without RAS and with and without recycling agents. Asphalt pavement distresses such as permanent deformation, terminal IRI, fatigue cracking, and raveling were predicted for three different pavement designs, two climatic conditions, and five asphalt mixtures. The three pavement designs selected for this study (Fig. 1), represent common asphalt pavement structures used with three traffic volume levels (low, medium, and high). In order to compare the effects of RAS on pavement performance, the properties of the surface layer were changed while the properties of the other layers were kept constant [33]. 4.2. Asphalt mixture description Table 1 summarizes the five asphalt mixtures considered in this study. The recycling agents evaluated in this study were a
Three different mechanistic tests (i.e., dynamic complex modulus, Semi-Circular Bending [SCB] test and Loaded Wheel Tracking [LWT]) were conducted on the five mixtures (Table 3). These tests were conducted in order to evaluate the performance of each mixture against the different pavement distresses and to compare the results to the predicted performance from Pavement ME. 4.4. Design inputs The pavement structure was introduced to Pavement ME as a new flexible pavement with a design life of 20 years. Since the mixes were compared relatively, default calibration factors were used in the analysis. A reliability level of 50% was assumed in Pavement ME since the mixes were compared relatively [33]. 4.4.1. Traffic Traffic data consisted of the Average Annual Daily Traffic (AADT) values for multiple traffic classifications, truck factors, together with the distribution for vehicle classes 1–13. A truck
Fig. 1. Pavement designs for (a) high volume traffic, (b) medium volume traffic, and (c) low volume traffic.
690
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
Table 1 Asphalt mixtures properties. Mixture ID
Description
70CO 70PG5P 70PG5M 70PG5P5HG
Virgin Binder Virgin Binder Virgin Binder Virgin Binder 5% RA1 Virgin Binder 12% RA1
70PG5P12CYCL
Shingle type
RA additive
70-22 70-22 + 5%RAS 70-22 + 5%RAS 70-22 + 5%RAS +
N/A PCWS MWS PCWS
N/A N/A N/A RA2
PG 70-22 + 5%RAS +
PCWS
RA1
PG PG PG PG
traffic classification (TTC) of nine was assumed in the analysis; previous research provided the monthly distribution for the selected TTC [34]. The hourly distribution factor and growth factor were kept as the default values in Pavement ME. The values of the AADT were 816, 1992, and 14554 for low, medium, and high traffic volumes, respectively.
had the highest |E⁄| values at high frequency, while 70PG5P had the lowest |E⁄| values. High stiffness of 70PG5P12CYCL was due to the high contribution of RAS binder when the recycling agent was used. 4.4.4. Base and subgrade properties As shown in Fig. 1, the selected base and subgrade were crushed limestone and clayey subgrade, respectively. The values of the resilient modulus (Mr) for these materials were acquired from previous projects [35]. Base and subgrade materials were the same for all three-pavement designs; therefore, these values were kept constant. The resilient modulus was 28.6 ksi for crushed limestone base and 10.4 ksi for clayey subgrade.
4.4.2. Climate Two different climatic conditions were considered from the Pavement ME climate database to represent cold and hot climatic conditions in the United States. The selected locations were Champaign, IL and Baton Rouge, LA for cold and hot climates, respectively.
4.4.5. Cost effectiveness A simplified Life-Cycle Cost (LCC) was conducted to evaluate the cost effectiveness of RAS in asphalt pavements. The simplified LCC was calculated as the overall production cost of one ton of each mixture, and divided by the predicted service life (years), which was obtained for three traffic levels by Pavement ME [33]. The RAS cost assumed in this study was based only on the material cost and not including the fraction of the transportation cost given its high dependency on the distance between plant and construction site. Table 4 summarizes the cost values obtained from contractors and recycling agents’ producers for each material type.
4.4.3. AC layer properties Mixture properties were introduced to Pavement ME as a level 1 input by measuring the dynamic complex modulus (E⁄) and binder complex shear modulus (G⁄), in accordance with AASHTO T 342. Dynamic moduli were measured using triplicate samples with dimensions of 150 mm in height and 100 mm in diameter. Testing was performed at loading frequencies of 0.1, 0.5, 1.0, 5, 10 and 25 Hz at temperatures of 10, 4, 20, 38.8 and 54.4 °C. A comparison of the master curve constructed for each mixture is presented in Fig. 2. It is noted that the 70PG5P12CYCL mixture
4.4.6. Performance indicators After defining the inputs into Pavement ME, pavement performance was predicted for the different designs, traffic loading levels, and mixes. Pavement performance was predicted using structural response models (mechanistic models) and transfer functions (empirical models) incorporated in the design software. For flexible pavements, top-down cracking, bottom-up cracking, thermal cracking, roughness, and rutting were the considered performance indicators [36]. The bottom-up fatigue cracking performance prediction model in Pavement ME is defined as follows [37]:
Table 2 Job-mix formula. Property
70CO
70PG5M
70PG5P
70PG5P5HG
70PG5P12CYCL
%Gmm at Ninitial %Gmm at Nfinal Air Voids (%) VMA (%) VFA (%) Total AC (%) Virgin AC (%) AC from RAS (%) Recycled Binder Ratio (%)
88.8 97.0 4.0 13.3 70 5.3 5.3 0 0
89.0 97.1 4.0 13.8 71 5.3 4.7 0.6 11.3%
88.9 96.9 4.0 13.9 71 5.3 4.8 0.5 9.4%
89.4 96.9 4.0 13.3 69 5.3 3.9 1.4 26%
88.9 96.5 4.5 13.6 67 5.3 3.9 1.4 26%
100 96 84 62 44 31 23 16 9 5.3 94.7 2.506 2.561 0.9 4.5 1.2 Pass
100 97 86 64 45 32 24 17 9 5.2 94.7 2.514 2.568 0.9 4.5 1.1 Pass
100 97 85 63 45 32 24 16 9 5.2 94.7 2.514 2.598 1.3 4 1.3 Pass
100 97 86 63 45 32 24 16 9 5.2 94.9 2.514 2.585 1.1 4 1.3 Pass
Metric (U.S.) Sieve
Gradation
19 mm (¾in) 12.5 mm (½in) 9.5 mm (⅜in) 4.75 mm (No.4) 2.36 mm (No.8) 1.18 mm (No.16) 0.6 mm (No.30) 0.3 mm (No.50) 0.15 mm (No.100) 0.075 mm (No.200) Ps, % Gsb Gse Pba, % Pbe, % D/Pbe D/Pbe, Spec 0.6–1.6
100 97 85 63 44 32 24 17 8 5.3 94.7 2.522 2.582 0.9 4.4 1.2 Pass
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697 Table 3 Mixture performance tests. Test *
Dynamic Modulus, |E | LWT SCB
Nf ¼ k1 CC H with
C ¼ 10 CH ¼
Samples Dimension
AASHTO T 342 AASHTO 324 AASHTO TP 105
£150 mm 100 mm £150 mm 60 mm £150 mm 57 mm
k2 k3 1 1 jE j et
4:84
Protocol
ð1Þ
V be 0:69 V a þV be
1 0:000398 þ 1þe0:003602 11:023:49HAC
ð2Þ
where, Nf = allowable number of load repetitions to fatigue cracking; et = tensile strain at the bottom of asphalt layers beneath the wheel; |E⁄| = dynamic modulus of the asphalt mixture; CH = thickness correction term for fatigue cracking; k1–3 = global field calibration coefficients; Vbe = effective binder content by volume,%; Va = air void content,%; and HAC = total thickness of asphalt layers. 5. Results and analysis 5.1. Pavement ME performance prediction Pavement ME was used to predict asphalt pavement distresses for the three pavement designs under three traffic loading levels (low, medium, and high), as well as for two climatic conditions (cold and hot regions). In order to be able to evaluate the effect of RAS on pavement performance, HMA layer properties were changed in the design, while other layer properties were kept
691
constant. The relative IRI, relative rutting, relative fatigue cracking, and relative service life predicted by Pavement ME was calculated in reference to the performance of the conventional mixture without RAS and recycling agents (i.e., 70CO). Relative values can be defined as the ratio of IRI, rutting, fatigue cracking and service life of the mixtures (70PG5P, 70PG5M, 70PG5PHD and 70PG5PCYCL) divided by the correspondent values of the control mixture (70CO). Furthermore, the cost of asphalt mixtures containing RAS with or without recycling agents was compared to the cost of the conventional asphalt mixture to evaluate the cost effectiveness of sustainable mixtures. 5.1.1. International roughness index (IRI) Fig. 3 illustrates the effects of RAS on the terminal IRI after 20 years in service. As shown in Fig. 3(a) and (b), the best IRI performance prediction, when compared to the control mixture, was for the mixture containing 5% RAS (PCWS) and 5% RA2 (70PG5P5HG), in contrast to the mixture containing 5% RAS and no RA (70PG5P), which showed the worst roughness performance. As shown in Fig. 3, the ranking was identical for both the cold and hot climates. 5.1.2. Permanent deformation in HMA layer The relative rut depth of the mixtures with and without RAS and RAs are presented in Fig. 4. The 70PG5P5HG mixture (mix with 5% PCWS and 5% RA2) showed the best rutting performance at the three traffic levels. Yet, the rutting performance was very close for the rest of the mixtures as indicated by a relative rut depth close to one. A Hamburg Loaded-Wheel Test (LWT) was conducted in accordance to AASHTO T 324 to measure the rut depth of the five mixtures. The test was stopped when the specimen had completed 20,000 cycles or reached a 20 mm deformation. A rut depth lower than 6 mm after 20,000 cycles is desirable [28]. As shown in Fig. 4 (c), 70PG5M and 70PG5P5HG mixtures had the best rutting performance as compared to the conventional mix (70CO). Comparison of the results obtained from Pavement ME and LWT showed a good agreement in ranking for hot climate; however, a slight discrepancy was observed in cold climate. This could be related to the
Fig. 2. Master curves of asphalt mixes.
692
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
where,
Table 4 Paving materials costs.
1 2
Item
Cost
Asphalt Cement Virgin Aggregate PCWS1 & MWS2 Aggregate Rejuvenator Additive (RA1) Rejuvenator Additive (RA2)
$475/Liquid Ton $35/Ton $27/Ton $1960/Ton $2080/Ton
Jc = critical strain energy release rate (kJ/mm2); b = sample thickness (mm); a = notch depth (mm); U = strain energy to failure (N.mm); and dU/da = change of strain energy with notch depth.
PCW: Post-Consumer Waste Shingle. MWS: Manufacture Waste Shingle.
LWT test conditions, which simulate high temperature service conditions. 5.1.3. AC Bottom-up fatigue cracking Fig. 5(a and b) shows the effect of RAS on bottom-up fatigue cracking. Similar to rutting performance, 70PG5P5HG mix (mix with 5% PCWS and 5% RA2) showed the best performance, as shown in Fig. 5. The worst performer, was also the 70PG5P mix, which was the mix with 5% RAS and no RA. The mixture containing RA1 showed good fatigue performance especially at medium and high traffic levels. A semi-circular bending (SCB) test was conducted to evaluate the laboratory fracture resistance of the mixtures at intermediate temperature. In this test, semi-circular specimens with three different notch depths (25.4, 31.8 and 38.1 mm) were tested at 25 °C to evaluate the change of strain energy with notch depth (dU/da). Subsequently, the Jc value was computed by dividing dU/da by the specimen thickness:
Jc ¼
1 dU b da
ð3Þ
A typical failure criterion for the Jc is a minimum of 0.50 kJ/m2 [38]. Fig. 5c, presents the critical strain energy release rate (Jc) for the five asphalt mixtures evaluated in this study. As shown in Fig. 5c, the mixtures 70PG5P5HG and 70PG5P12CYCL were the worst performers with a Jc of 0.3 and 0.4 kJ/m2, respectively, which were lower than the minimum failure criterion. This can be explained by the increased availability of aged RAS binder in these mixes given the action of the RA. It is noted that these mixtures had the highest RBRs. Based on these results, there is an inconsistency in the prediction between Pavement ME and SCB test results since the 70PG5P5HG mixture had the best performance in Pavement ME and the worse performance in the SCB test. Based on the results presented in Fig. 5, contradicting predictions were observed between Pavement ME and the SCB test results. This is due to the stiffening effect of RAS, which is reflected in the dynamic modulus inputted in Pavement ME, and results in stiffer mixes showing better performance. To address this limitation in Pavement ME, the performance model for fatigue cracking should be modified to consider other mixture characteristics in addition to the dynamic modulus such that it can successfully reflect the fracture resistance of different classes of asphalt mixtures including the ones incorporating RAS. Based on accelerated pavement testing of different classes of asphalt mixtures, the authors have proposed a modified AASHTO model for fatigue cracking, which incorporates the critical strain energy release rate
Fig. 3. Relative terminal IRI comparison (a) cold region and (b) hot region.
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
693
Fig. 4. Relative permanent deformation comparison: (a) cold region, (b) hot region, and (c) LWT lab results.
(Jc) measured from SCB. The original fatigue model presented in Eq. (1) was modified according to the following equation [39]:
Nf ¼ k1 CC H
k2 k3 1 1 ðJ c Þk4 jE j et
ð4Þ
where, k4 = the exponent of Jc, and all the other parameters and coefficients have been defined in Eq. (1). Results showed that the modified model produced a more realistic correlation with the SCB test results and field performance [39]. 5.1.4. AC Top-down fatigue cracking Fig. 6(a and b) shows the effect of RAS on top-down fatigue cracking. Similar to bottom-up cracking performance, 70PG5P5HG mix (mix with 5% PCWS and 5% RA2) showed the best performance. The worst performer, was also the 70PG5P mix, which was the mix with 5% RAS and no RA. This trend was expected as Pavement ME deals with top-down cracking similar to bottomup fatigue cracking. 5.1.5. Service life prediction The service life of the studied cases was defined as the time at which the pavement design exceeds the terminal threshold set
for the different performance distresses. In almost all cases, the critical distress was the HMA rutting performance criterion. Fig. 7 shows the relative design life as compared to the conventional mix (70CO). As reflected in the previous discussions, mix 70PG5P5HG outperformed all the other mixes considered in the analysis. 5.2. Statistical analysis A statistical analysis was performed to assess the overall performance of the mixtures for the different combinations possible between the analyzed climate conditions and traffic levels (i.e., six iterations per mixture). An Analysis of Variance (ANOVA) was performed for each pavement distress to determine whether there was statistical difference in the mixture performance. Results of the ANOVA showed that there were significant differences between the mixture performances at a confidence level of 0.05. Therefore, a t-test was performed at a confidence level of 0.05 on all the possible combinations (i.e., 70CO vs. 70PG5P vs. 70PG5M vs. 70PG5HG) to identify the mixes that were statistically different. Table 4 shows the statistical ranking of the mixtures for the different pavement distresses. The statistical results of each grouping is ranked by using letters A, B, C, and so forth. The letter A is assigned
694
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
Fig. 5. Relative bottom-up fatigue cracking comparison: (a) cold region, (b) hot region, and (c) SCB test results.
to the mix with the highest mean followed by the letter B and so forth. Double letters (e.g., A/B, B/C) indicate that the mixture may be categorized in both groups. Results presented in Table 5 shows that the best mixture overall against roughness was 70PG5P5HG mix, which was statistically different from the rest of the mixtures. Similarly, the 70PG5P5HG mix was the best performer against rutting while mix 70CO was the worst performer in the different combinations available between the two climatic conditions and the three traffic levels evaluated in the study. A similar trend was observed for the bottom-up and top-down fatigue cracking where the best performance mixture (70PG5P5HG) was statistically different from the worst performance mixture (70PG5P). Based on the statistical results, it can be observed that the addition of 5% RA2 improved the performance of a mixture containing oxidized asphalt binder (i.e., 5% RAS) compared to the mixture with only RAS (i.e., 70PG5P and 70PG5M) and the mixture with pure virgin binder (70CO). However, the predicted performance of Pavement ME differed from the predicted SCB test results since Pavement ME does not incorporate mix resistance to cracking in the design.
5.3. Cost and energy analysis Fig. 8(a) shows the production cost of the five different mixtures. Mixtures with RAS are more economical to produce. Yet, the cost of 70PG5PCYCL was much higher due to the higher percentage of rejuvenator additive relative to the other mixtures. On the other hand, 70PG5M was the cheapest mixture to produce. The production cost was divided by the predicted service life in order to determine the mixtures that would be the most cost effective over the life cycle of the pavement. As shown in Fig. 8(b and c), 70PG5P5HG (mix with 5% PCWS and 5% RA2) had the lowest cost over the life cycle for every traffic condition and for both climatic regions. 6. Summary and conclusions This study had two main objectives: (1) to evaluate the effects of RAS (with and without recycling agents) on the predicted performance from Pavement ME; and (2) to assess the life cycle costs of pavements constructed with RAS and recycling agents.
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
695
Fig. 6. Relative top-down fatigue cracking comparison: (a) cold region and (b) hot region.
Fig. 7. Relative pavement service life: (a) cold region and (b) hot region.
An evaluation of the effects of RAS and RAs was conducted using Pavement ME for two different climatic conditions (cold and hot regions).
Based on the results of the analysis, Pavement ME predicted that the mix with 5% PCWS and 5% RA2 (70PG5P5HG) will be the best performer against roughness, HMA rutting, and fatigue
696
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
Table 5 Statistical ranking of mixtures. Mixture type
70CO 70PG5P 70PG5M 70PG5P5HG 70PG5P12CYCL
Terminal IRI
AC permanent Deformation
Bottom-up Fatigue cracking
Top-down Fatigue cracking
Mean
Rank
Mean
Rank
Mean
Rank
Mean
Rank
1.000 1.003 0.990 0.961 0.996
B A C D C
1.000 0.996 0.946 0.784 0.989
A B D E C
1.000 1.020 0.988 0.955 0.984
B A C D C/D
1.000 1.040 0.978 0.906 0.970
B A C D C/D
Fig. 8. Cost-effectiveness of mixes with RAS: (a) production cost comparison, (b) cold region, and (c) hot region.
cracking. This is due to the stiffening effect of RAS, which is reflected in the dynamic modulus inputted in the software, and results in stiffer mixes showing better performance. While one would expect RAS to improve rutting performance, the superior fatigue performance was not expected given that the binder in RAS is an air-blown asphalt binder with poor elongation and relaxation characteristics. Results of SCB test results realistically predicted that mixes with RAS and recycling agents would be the worst performers against cracking. This can be explained by the increased availability of aged RAS binder in these mixes when a
recycling agent is used. As presented in this study, the fatigue performance model in Pavement ME should be modified to consider other mixture characteristics in addition to the dynamic modulus such that it would reflect the fracture resistance of different classes of asphalt mixtures including the ones incorporating RAS. Results of the cost analysis showed that the mixtures with RAS are more economical to produce. Yet, the cost of 70PG5PCYCL was much higher due to the higher percentage of rejuvenator additive relative to the other mixtures. On the other hand, the mix with RAS and no recycling agent was the cheapest mixture to produce. The
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
production cost was divided by the predicted service life in order to determine the mixtures that would be the most cost effective over the life cycle of the pavement. When considering the predicted performance of the mixes, the mix with 5% PCWS and 5% RA2 had the lower cost over the life cycle for every traffic condition and for both climatic regions. References [1] L. Uzarowski, Sustainable Pavements – Making the Case for Longer Design Lives for Flexible Pavements, Annual Conference of the Transportation Association of Canada, Toronto, Ontario, 2008. [2] N.H. Tran, A. Taylor, R. Willis, Effect of Rejuvenator on Performance Properties of HMA Mixtures with High RAP and RAS Contents Report 12-05, National Center for Asphalt Technology, Auburn University, Auburn, Ala, 2012. [3] R. Karlsson, U. Isacsson, Material-related aspects of asphalt recycling—state of the art, J. Mater. Civil Eng. 18 (2006). [4] E. Brown, Preventative Maintenance of Asphalt Concrete Pavements, National Center for Asphalt Technology, Auburn University, Alabama, 1988. [5] Y.H. Huang, Pavement analysis and design, Guide for Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures: Part 1. NCHRP 1-37A Final Report, second ed., Pearson Prentice Hall, 2004. [6] A. Buss, M. Rashwan, T. Breakah, R.C. Williams, A. Kvasnak, Investigation of Warm-Mix Asphalt Using the Mechanistic-Empirical Pavement Design Guide, Mid-Continent Transportation Research Symposium, Ames, IA, 2009. [7] S.B. Cooper Jr., Characterization of HMA Mixtures Containing High Recycled Asphalt Pavement Content with Crumb Rubber Additives Thesis (MSCE), State University, 2008. [8] J.S. Daniel, G.R. Chehab, D. Ayyala, Sensitivity of RAP binder grade on performance predictions in the MEPDG, Annual Meeting and Technical Sessions CD, Association of Asphalt Paving Technologists (AAPT), Minneapolis, MN, 2009. [9] S. Diefenderfer, A. Hearon, Laboratory Evaluation of a Warm Asphalt Technology for Use in Virginia VTRC 09-R11 Final Report, Virginia Transportation Research Council, Charlottesville, VA, 2008. [10] S.W. Goh, Y. Zhanping, T.J. Van Dam, Laboratory evaluation and pavement design for warm mix asphalt, Mid-Continent Transportation Research Symposium, 2007. [11] M. Hassan, Life-cycle assessment of warm-mix asphalt: an environmental and economic perspective, Transportation Research Board 88th Annual Meeting, 2009. [12] G.W. Maupin, S.D. Diefenderfer, J.S. Gillespie, Performance and economic evaluation of virginia’s higher RAP specification, in: Transportation Research Record: Journal of the Transportation Research Board, No. 2126, 2009, pp. 142–150. [13] L.M. Mohammad, M.A. Elseifi, S.B. Cooper III, Laboratory evaluation of sulfurextended asphalt mixture, Second International Conference on Construction In Developing Countries, 2010. Cairo, Egypt. [14] S.T. Muench, Oadway construction sustainability impacts: a review of life cycle assessments, Transportation Research Board 89th Annual Meeting, 2010. [15] D. Timm, N. Tran, A. Taylor, M. Robbins, B. Powell, Evaluation of Mixture Performance and Structural Capacity of Pavements Utilizing Shell ThiopaveÒ Phase I: Mix Design, Laboratory Performance Evaluation and Structural Pavement Analysis and Design NCAT Report 09-05, National Center for Asphalt Technology, Auburn, AL, 2009. [16] National Cooperative Highway Research Program (NCHRP), Guide for Mechanistic-Empirical Design of New and Rehabilitated Structures NCHRP Report 01-37A, Transportation Research Board, Washington, DC, 2004. [17] B. Nam, H. Mahernia, Behzadan, H. Amir, Mechanical characterization of asphalt tear-off roofing shingles in Hot Mix Asphalt, Constr. Build. Mater. 50 (2014) 308–316. [18] J.R. Willis, P. Turner, Characterization of Asphalt Binder Extracted From Reclaimed Asphalt Shingles, National Center for Asphalt Technology, 2016. [19] Northeast Recycling Council, Inc., Asphalt Shingles Waste Management in the northeast Fact Sheet, 2011. http://www.nerc.oorg/documents/asphalt.pdf.
697
[20] N. Ali, J.S. Chan, A. Potyondy, R. Bushman, A. Bergen, Mechanistic evaluation of asphalt concrete mixtures containing reclaimed roofing materials, Proceedings of the 74thAnnual Meeting of Transportation Research Board, ISSN: 03611981, 1995. [21] A. Alvergue, M. Elseifi, L. Mohammad, S.B. Cooper, S. Cooper, Laboratory evaluation of asphalt mixtures with reclaimed asphalt shingle prepared using the wet process, Road Mater. Pavement Des. 15 (2014) 62–77. [22] D.E. Newcomb, M.S. Gardiner, B.M. Weikle, A. Drescher, Influence of roofing shingles and asphalt concrete mixture properties, Department of Civil and Mineral Engineering, University of Minnesota, 1993. [23] K.Y. Foo, D.I. Hanson, T.A. Lynn, Valuation of roofing shingles in hot mix asphalt, J. Mater. Civil Eng. (1999). [24] B. Sengoz, A. Topal, Use of asphalt roofing shingle waste in HMA, Constr. Build. Mater. 19 (2005) 337–346. [25] S. Tighe, N. Hanasoge, B. Eyers, R. Essex, S. Damp, Who Thought Recycled Asphalt Shingles (RAS) Needed To Be Landfilled: Why Not Build a Road?, Recycled Materials and Recycling Processes for Sustainable Infrastructure Session (Annual Conference of the Transportation Association of Canada), 2008 [26] K. Anurag, F. Xiao, S.N. Amirkhanian, Laboratory investigation of indirect tensile strength using roofing polyester waste fibers in hot mix asphalt, Constr. Build. Mater. 23 (5) (2009) 2035–2040. [27] G.W. Maupin, Investigation of the Use of Tear-Off Shingles in Asphalt Concrete, Virginia Transportation Research Council, 2010. [28] Samuel B. Cooper, Sustainable Materials for Pavement Infrastructure: Design and Performance of Asphalt Mixtures Containing Recycled Asphalt Shingles, Dissertation Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College, Baton Rouge, LA, 2015. [29] C. Robinette, J. Epps, Energy, emissions, material conservation, and prices associated with construction, rehabilitation, and material alternatives for flexible pavement, J. Transp. Res. Board (2010) 2179. [30] J.W. Arnold, B. Behnia, M.E. McGovern, B. Hill, W.G. Buttlar, H. Reis, Quantitative evaluation of low-temperature performance of sustainable asphalt pavements containing recycled asphalt shingles (RAS), Constr. Build. Mater. 58 (2014) 1–8. [31] A. Cascione, R.C. Williams, J. Yu, Performance testing of asphalt pavements with recycled asphalt shingles from multiple field trials, Constr. Build. Mater. 101 (2015) 628–642. [32] R. Ghabchi, M. Barman, D. Singh, M. Zaman, M.A. Mubaraki, Comparison of laboratory performance of asphalt mixes containing different proportions of RAS and RAP, Constr. Build. Mater. 124 (2016) 343–351. [33] Samuel B. Cooper, M. Elseifi, L. Mohammed, M. Hassan, Performance and cost effectivenes of sustainable pavement technologies in flexible pavement using the mechanistic-empirical pavement design guide, Am. Soc. Civil Eng. J. (2013). [34] S. Ishak, H. Shin, B. Sridhar, Characterization and development of truck load spectra and growth factors for current and future pavement design practices in louisiana LTRC Report 07-2P, Louisiana Transportation Research Center, Baton Rouge, LA, 2009. [35] L.M. Mohammad, A. Herath, R. Gudishala, M.D. Nazzal, M.Y. Abu-Farsakh, K. Alshibli, Development of Models to Estimate the Subgrade and Subbase Layers’ Resilient Modulus from In situ Devices Test Results for Construction Control LTRC Report 02-4B, Louisiana Transportation Research Center, Baton Rouge, LA, 2008. [36] R.L. Baus, N.R. Stires, Mechanistic-Empirical Pavement Design Guide Implementation, 2010. Report No. FHWA-SC-10-01, FHWA/SCDOT. [37] AASHTO – American Association of State Highway and Transportation Officials, Mechanistic-Empirical Pavement Design Guide – A manual of Practice, second ed., Washington D.C., 2015. [38] Z. Wu, L. Mohammad, B. Wang, M.A. Mull, Fracture resistance characterization of superpave mixture using the semi-circular bending test, J. ASTM Int. 2 (3) (2005) 1–15. [39] W. Cao, L.N. Mohammad, M.A. Elseifi, Asphalt pavement fatigue performance prediction based on semicircular bending test at intermediate temperature, Paper No. 17-05160 Presented at the 96th Transportation Research Board Annual Meeting, Washington, D.C, 2017.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Mechanistic-empirical pavement performance of asphalt mixtures with recycled asphalt shingles Sharareh Shirzad a, Max A. Aguirre a, Luis Bonilla a, Mostafa A. Elseifi b,⇑, Samuel Cooper c, Louay N. Mohammad b a b c
Louisiana State University, United States Department of Civil and Environmental Engineering, Louisiana State University, United States Louisiana Transportation Research Center, United States
h i g h l i g h t s The effect of recycled asphalt shingles on pavement performance was evaluated. Pavement Mechanistic-Empirical (ME) predicted rutting performance correctly. There was an inconsistency in fatigue cracking prediction for mixes with RAS.
a r t i c l e
i n f o
Article history: Received 26 September 2017 Received in revised form 17 November 2017 Accepted 20 November 2017 Available online 24 November 2017 Keywords: Asphalt mixture Recycled asphalt shingle Mechanistic-empirical pavement design Recycling agents Semi-circular bending
a b s t r a c t As highway agencies are in the process of adopting the new mechanistic-empirical pavement design guide, Pavement ME, it is unclear how asphalt mixtures incorporating Recycled Asphalt Shingle (RAS) will influence the design when mechanistic-empirical approaches are used. In this study, Pavement ME was used to evaluate the effects of RAS (with or without recycling agents [RAs]) on pavement performance. Furthermore, a cost analysis was conducted to assess the life-cycle cost of asphalt pavements constructed with RAS. Three different pavement structures were analyzed at three traffic levels (low, medium, and high) and for two climatic regions (cold and hot). Pavement ME predicted that the mix with 5% PostConsumer Waste Shingle (PCWS) and 5% RA would be the best performer against roughness, rutting, and fatigue cracking. This is due to the stiffening effect of RAS, which is reflected in the dynamic modulus inputted in the software. While one would expect RAS to improve rutting performance, the superior fatigue performance of mixes with RAS was not expected given that the binder in RAS is an air-blown asphalt binder with poor elongation and relaxation characteristics. Results were compared to the Semi-Circular Bending (SCB) test results, which realistically predicted that the mixes with RAS and recycling agents would be the worst performers against cracking. This can be explained by the increased availability of aged RAS binder in these mixes when a recycling agent is used. Results of the cost analysis showed that the mixtures with RAS are more economical to produce. When considering the predicted performance of the mixes, the mix with 5% PCWS and 5% RA had the lowest cost over the pavement service life. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction The sustainability movement in road applications aims to achieve production, distribution, and construction of asphalt mixtures designed to last longer with less impact on ecological systems. Sustainable pavements would also minimize the use of natural resources, reduce energy consumption and greenhouse gas emissions, limit pollution, improve health and safety, and ⇑ Corresponding author. E-mail address: [email protected] (M.A. Elseifi). https://doi.org/10.1016/j.conbuildmat.2017.11.114 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
ensure a high level of user comfort [1]. Utilizing recycled materials to decrease the amount of virgin asphalt binder or aggregate used in the production of asphalt mixtures is an example of sustainable pavement technologies. Studies have shown that using recycled materials, such as recycled asphalt shingles (RAS), reduces the environmental impacts related to extraction, transportation, and processing of virgin materials [2]. Different recycling agents (RAs) may also be used in the asphalt mixtures containing RAS to improve the blending between recycled and virgin materials [3]. Recycling agents may also have the ability to reverse the aging process by softening or
688
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
rejuvenating the aged binder [4]. In addition to the environmental benefits, cost reduction associated with the use of RAS as well as the cost increase due to the use of recycling agents are important factors that should be evaluated when studying asphalt mixtures incorporating RAS and RAs. Traditionally, pavements were designed based on empirical principles, which used road tests results conducted by the American Association of State Highway Officials (AASHO) in the late 1950s and early 1960s in Ottawa, Illinois. The limitations of empirical design include that: (1) the design considers pavement performance through a subjective rating; (2) it offers limited attention to failure modes; and (3) it depends on conditions remaining the same, similar to the original field conditions. Yet, a 54% increase in the number of vehicles using the roads and a 75% increase in the vehicle miles traveled was observed between 1973 and 1993; thereby making today’s traffic loads significantly higher than the ones utilized to develop the semi-empirical 1993 AASHTO design method [5]. Based on these limitations, a Mechanistic-Empirical (ME) design method was developed, which accounts for variability in material properties, traffic loads, and construction procedures. In this new approach, both mechanical and empirical methods are incorporated by calculating pavement responses due to loading then relating those responses to pavement performance. Previous studies have evaluated the mechanistic properties, economic benefits, and ecological impacts from sustainable pavement technologies [6–15]. Yet, the incorporation of RAS and RAs in pavement design has not been thoroughly evaluated using ME design approaches. 2. Objectives and scope This study had two main objectives: (1) to evaluate the effects of RAS (with and without recycling agents) on the predicted performance from Pavement ME; and (2) to assess the life cycle cost of pavement constructed with RAS and recycling agents. An evaluation of the effects of RAS and RAs was conducted using Pavement ME for two different climatic conditions (cold and hot regions). Furthermore, the analysis encompassed three traffic levels (low, medium, and high) to evaluate the effect of traffic load on the performance of RAS materials. Level 1 inputs were used to describe the asphalt layers while Level 2 inputs were used to describe the granular base and subgrade layers. 3. Background New methodology, developed as part of NCHRP Project 1-37A, (Development of the 2002 AASHTO Guide for Design of New and Rehabilitated Pavement Structures: Phase II), sought to relate pavement distresses to pavement responses (stress, strain and deflection); this process is based on traffic loading, climate data, and material properties. These distresses may be used to predict pavement damage over time. Pavement ME is based on an iterative process, which initiates the analysis with a trial design and then evaluates the distress susceptibility of the primary design. If the predicted distresses meet the specified performance criteria, the primary design is accepted; otherwise, the evaluation process is repeated with a revised design. To utilize available data in the most efficient way, three different levels of input are considered in Pavement ME. This approach provides the user with the ability to set the level of input data according to pavement importance. Level 1 inputs evaluate pavement performance with respect to the mechanistic properties of the different layers, which in turn can be measured through laboratory or field testing. Therefore, Level 1 input has the highest accuracy, coupled with the highest expenses. Level 2 input addresses those predicted mechanistic val-
ues, which may be obtained from different databases or extrapolated from less extensive tests. Level 3 input considers the typical default values and therefore displays the lowest accuracy. Interestingly, the computational methodology used for distress prediction is the same for all input levels [16]. 3.1. Recycled asphalt shingles Recycled asphalt shingles, which represent one of the largest municipal solid wastes, can be used as a virgin binder replacement in asphalt pavements [17]. Every year, 11 million tons of waste shingles are produced, which results in 22 million cubic yards of waste materials, which must be landfilled [18]. Nineteen to 36% of RAS consist of a relatively hard asphalt binder; 2–15% fiberglass or cellulose backing, 20–38% fine aggregate, ceramic-coated natural rock, and the remainder of 1–12% is a mineral filler or stabilizer [19]. Reusing this waste material in asphalt pavement construction could present significant environmental and economic benefits. Starting from the 1980s, researchers evaluated the use of RAS in HMA production. Several studies were conducted to evaluate the impact of RAS on asphalt mixture performance [20–25]. Anurag et al. utilized the indirect tensile strength test to determine the effect of roofing waste with polyester fibers with different fiber lengths and different fiber contents on the moisture sensitivity of the mix. The study concluded that polyester fibers could improve the wet tensile strength and tensile strength ratio of the mixture [26]. More recently, Maupin used 5% RAS in asphalt mixture to evaluate the effect of RAS on the mix durability [27]. Results from the extracted binder material showed that the high-temperature grading of the binder was improved, while the low-temperature grading was not affected. Maupin also used mixture laboratory testing to evaluate the rutting, cracking, and tensile strength of the asphalt mixtures. Results of rut testing indicated that the mixes would perform satisfactorily on high traffic conditions. Similarly, the mixes were expected to perform satisfactorily against fatigue failure. Cooper examined the performance of different mixtures containing RAS with and without recycling agents [28]. The tests encompassed laboratory mechanistic tests to characterize the low, intermediate, and high temperature performance of mix, including the dynamic modulus test, semi-circular bending (SCB) test for intermediate temperature, thermal stress restrained specimen tensile strength test (TSRST) for low temperature, and a Hamburg type Loaded Wheel Tracking (LWT) test to evaluate rutting resistance. Results showed that although RAS caused a reduction in virgin binder, the extracted binder did not blend completely with the virgin binder. While there was no improvement in intermediate temperature cracking, the rutting resistance was improved [28]. Robinette and Epps studied the influence of RAS on energy consumption, emissions generation, and natural resource consumption [29]. They also reviewed the price of construction for mixtures containing RAS as compared to conventional materials and construction. Results indicated that in most cases, RAS addition could reduce energy consumption, generated emissions, and preserve natural resources such as aggregate and asphalt binder. By virtue of these processes, RAS can reduce the price of asphalt mixture production and construction as well as alleviating the overall environmental impacts. Arnold et al. examined the low-temperature cracking characteristics of asphalt mixtures containing various amount of RAS (0– 12.5%) using the Disc-shaped Compact Tension (DCT) and acoustic emission (AE) tests. Result showed that the mixtures with RAS had lower fracture energies and higher peak loads. The increase in peak load with the increase in RAS content indicated that the use of RAS did not affect the HMA’s high-temperature performance. However,
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
it significantly affected the embrittlement temperature of the mixtures [30]. Cascione et al. tested permanent deformation, fatigue cracking, and low temperature cracking performance of asphalt mixtures with RAS from seven different transportation agencies. When RAS was added to the mixtures, the high and low temperature grades of the virgin binder were increased. Furthermore, based on the flow number and dynamic modulus results, it was concluded that RAS addition improved the rutting resistance of the mixtures. Performance of all the mixtures, with or without RAS, was acceptable with respect to fatigue cracking. Lowtemperature fracture energy was measured for SCB samples and results suggested that RAS could increase the fracture resistance of mixtures prior to long term aging [31]. In another attempt to evaluate the effect of recycled materials on the performance of asphalt mixtures, Ghabchi et al. [32] examined the fatigue and low-temperature cracking potential of mixtures with RAS and/or RAP. Results indicated that the addition of 5% RAS and 5% RAP to the mixtures with a straight asphalt binder led to a maximum increase in the fatigue life of the mixtures. However, increasing the amount of RAS beyond 5% caused a decrease in the fatigue life of the mixtures. The results obtained from the creep compliance test indicated that RAS addition increased lowtemperature cracking potential [32].
689
naphthenic oil (referred to as RA1) and a vegetable oil derived from the pyrolysis of pine tree (referred to as RA2). As shown in Table 1, the first mixture was the conventional mixture (70CO) containing no RAS or RAs. The second mixture (70PG5P) contained 5% PostConsumer Waste Shingle (PCWS) and no RAs. Similarly, the third mixture (70PG5M) contained 5% Manufacturer Waste Shingle (MWS) and no RAs. The fourth mixture (70PG5P5HG) contained 5% PCWS with 5% RA2. The fifth mixture (70PG5P12CYCL) contained 5% PCWS with 12% RA1. Optimum dosage of the recycling agents was selected by utilizing volumetric and densification criteria. The aggregate gradation and virgin asphalt cement percentage were kept approximately constant while the percentage of the recycling agent was increased. This was necessary for the comparison of the control mixture with mixtures containing recycling agents. Furthermore, the asphalt contribution by the recycled materials was estimated after identifying the optimum virgin asphalt content. Table 2 presents gradation and job mix formula of the mixtures evaluated in this study. It also demonstrates the amount of asphalt binder contribution from RAS and the recycled binder ratio (RBR) in mixtures 70PF5P, 70PG5M, 70PG5P5HG and 70PG5P12CYCL. More details on the selection of the RAs’ contents have been presented elsewhere [28]. 4.3. Experimental test matrix
4. Methodology 4.1. Pavement performance prediction In this study, Pavement ME was used to predict the performance of asphalt mixtures with and without RAS and with and without recycling agents. Asphalt pavement distresses such as permanent deformation, terminal IRI, fatigue cracking, and raveling were predicted for three different pavement designs, two climatic conditions, and five asphalt mixtures. The three pavement designs selected for this study (Fig. 1), represent common asphalt pavement structures used with three traffic volume levels (low, medium, and high). In order to compare the effects of RAS on pavement performance, the properties of the surface layer were changed while the properties of the other layers were kept constant [33]. 4.2. Asphalt mixture description Table 1 summarizes the five asphalt mixtures considered in this study. The recycling agents evaluated in this study were a
Three different mechanistic tests (i.e., dynamic complex modulus, Semi-Circular Bending [SCB] test and Loaded Wheel Tracking [LWT]) were conducted on the five mixtures (Table 3). These tests were conducted in order to evaluate the performance of each mixture against the different pavement distresses and to compare the results to the predicted performance from Pavement ME. 4.4. Design inputs The pavement structure was introduced to Pavement ME as a new flexible pavement with a design life of 20 years. Since the mixes were compared relatively, default calibration factors were used in the analysis. A reliability level of 50% was assumed in Pavement ME since the mixes were compared relatively [33]. 4.4.1. Traffic Traffic data consisted of the Average Annual Daily Traffic (AADT) values for multiple traffic classifications, truck factors, together with the distribution for vehicle classes 1–13. A truck
Fig. 1. Pavement designs for (a) high volume traffic, (b) medium volume traffic, and (c) low volume traffic.
690
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
Table 1 Asphalt mixtures properties. Mixture ID
Description
70CO 70PG5P 70PG5M 70PG5P5HG
Virgin Binder Virgin Binder Virgin Binder Virgin Binder 5% RA1 Virgin Binder 12% RA1
70PG5P12CYCL
Shingle type
RA additive
70-22 70-22 + 5%RAS 70-22 + 5%RAS 70-22 + 5%RAS +
N/A PCWS MWS PCWS
N/A N/A N/A RA2
PG 70-22 + 5%RAS +
PCWS
RA1
PG PG PG PG
traffic classification (TTC) of nine was assumed in the analysis; previous research provided the monthly distribution for the selected TTC [34]. The hourly distribution factor and growth factor were kept as the default values in Pavement ME. The values of the AADT were 816, 1992, and 14554 for low, medium, and high traffic volumes, respectively.
had the highest |E⁄| values at high frequency, while 70PG5P had the lowest |E⁄| values. High stiffness of 70PG5P12CYCL was due to the high contribution of RAS binder when the recycling agent was used. 4.4.4. Base and subgrade properties As shown in Fig. 1, the selected base and subgrade were crushed limestone and clayey subgrade, respectively. The values of the resilient modulus (Mr) for these materials were acquired from previous projects [35]. Base and subgrade materials were the same for all three-pavement designs; therefore, these values were kept constant. The resilient modulus was 28.6 ksi for crushed limestone base and 10.4 ksi for clayey subgrade.
4.4.2. Climate Two different climatic conditions were considered from the Pavement ME climate database to represent cold and hot climatic conditions in the United States. The selected locations were Champaign, IL and Baton Rouge, LA for cold and hot climates, respectively.
4.4.5. Cost effectiveness A simplified Life-Cycle Cost (LCC) was conducted to evaluate the cost effectiveness of RAS in asphalt pavements. The simplified LCC was calculated as the overall production cost of one ton of each mixture, and divided by the predicted service life (years), which was obtained for three traffic levels by Pavement ME [33]. The RAS cost assumed in this study was based only on the material cost and not including the fraction of the transportation cost given its high dependency on the distance between plant and construction site. Table 4 summarizes the cost values obtained from contractors and recycling agents’ producers for each material type.
4.4.3. AC layer properties Mixture properties were introduced to Pavement ME as a level 1 input by measuring the dynamic complex modulus (E⁄) and binder complex shear modulus (G⁄), in accordance with AASHTO T 342. Dynamic moduli were measured using triplicate samples with dimensions of 150 mm in height and 100 mm in diameter. Testing was performed at loading frequencies of 0.1, 0.5, 1.0, 5, 10 and 25 Hz at temperatures of 10, 4, 20, 38.8 and 54.4 °C. A comparison of the master curve constructed for each mixture is presented in Fig. 2. It is noted that the 70PG5P12CYCL mixture
4.4.6. Performance indicators After defining the inputs into Pavement ME, pavement performance was predicted for the different designs, traffic loading levels, and mixes. Pavement performance was predicted using structural response models (mechanistic models) and transfer functions (empirical models) incorporated in the design software. For flexible pavements, top-down cracking, bottom-up cracking, thermal cracking, roughness, and rutting were the considered performance indicators [36]. The bottom-up fatigue cracking performance prediction model in Pavement ME is defined as follows [37]:
Table 2 Job-mix formula. Property
70CO
70PG5M
70PG5P
70PG5P5HG
70PG5P12CYCL
%Gmm at Ninitial %Gmm at Nfinal Air Voids (%) VMA (%) VFA (%) Total AC (%) Virgin AC (%) AC from RAS (%) Recycled Binder Ratio (%)
88.8 97.0 4.0 13.3 70 5.3 5.3 0 0
89.0 97.1 4.0 13.8 71 5.3 4.7 0.6 11.3%
88.9 96.9 4.0 13.9 71 5.3 4.8 0.5 9.4%
89.4 96.9 4.0 13.3 69 5.3 3.9 1.4 26%
88.9 96.5 4.5 13.6 67 5.3 3.9 1.4 26%
100 96 84 62 44 31 23 16 9 5.3 94.7 2.506 2.561 0.9 4.5 1.2 Pass
100 97 86 64 45 32 24 17 9 5.2 94.7 2.514 2.568 0.9 4.5 1.1 Pass
100 97 85 63 45 32 24 16 9 5.2 94.7 2.514 2.598 1.3 4 1.3 Pass
100 97 86 63 45 32 24 16 9 5.2 94.9 2.514 2.585 1.1 4 1.3 Pass
Metric (U.S.) Sieve
Gradation
19 mm (¾in) 12.5 mm (½in) 9.5 mm (⅜in) 4.75 mm (No.4) 2.36 mm (No.8) 1.18 mm (No.16) 0.6 mm (No.30) 0.3 mm (No.50) 0.15 mm (No.100) 0.075 mm (No.200) Ps, % Gsb Gse Pba, % Pbe, % D/Pbe D/Pbe, Spec 0.6–1.6
100 97 85 63 44 32 24 17 8 5.3 94.7 2.522 2.582 0.9 4.4 1.2 Pass
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697 Table 3 Mixture performance tests. Test *
Dynamic Modulus, |E | LWT SCB
Nf ¼ k1 CC H with
C ¼ 10 CH ¼
Samples Dimension
AASHTO T 342 AASHTO 324 AASHTO TP 105
£150 mm 100 mm £150 mm 60 mm £150 mm 57 mm
k2 k3 1 1 jE j et
4:84
Protocol
ð1Þ
V be 0:69 V a þV be
1 0:000398 þ 1þe0:003602 11:023:49HAC
ð2Þ
where, Nf = allowable number of load repetitions to fatigue cracking; et = tensile strain at the bottom of asphalt layers beneath the wheel; |E⁄| = dynamic modulus of the asphalt mixture; CH = thickness correction term for fatigue cracking; k1–3 = global field calibration coefficients; Vbe = effective binder content by volume,%; Va = air void content,%; and HAC = total thickness of asphalt layers. 5. Results and analysis 5.1. Pavement ME performance prediction Pavement ME was used to predict asphalt pavement distresses for the three pavement designs under three traffic loading levels (low, medium, and high), as well as for two climatic conditions (cold and hot regions). In order to be able to evaluate the effect of RAS on pavement performance, HMA layer properties were changed in the design, while other layer properties were kept
691
constant. The relative IRI, relative rutting, relative fatigue cracking, and relative service life predicted by Pavement ME was calculated in reference to the performance of the conventional mixture without RAS and recycling agents (i.e., 70CO). Relative values can be defined as the ratio of IRI, rutting, fatigue cracking and service life of the mixtures (70PG5P, 70PG5M, 70PG5PHD and 70PG5PCYCL) divided by the correspondent values of the control mixture (70CO). Furthermore, the cost of asphalt mixtures containing RAS with or without recycling agents was compared to the cost of the conventional asphalt mixture to evaluate the cost effectiveness of sustainable mixtures. 5.1.1. International roughness index (IRI) Fig. 3 illustrates the effects of RAS on the terminal IRI after 20 years in service. As shown in Fig. 3(a) and (b), the best IRI performance prediction, when compared to the control mixture, was for the mixture containing 5% RAS (PCWS) and 5% RA2 (70PG5P5HG), in contrast to the mixture containing 5% RAS and no RA (70PG5P), which showed the worst roughness performance. As shown in Fig. 3, the ranking was identical for both the cold and hot climates. 5.1.2. Permanent deformation in HMA layer The relative rut depth of the mixtures with and without RAS and RAs are presented in Fig. 4. The 70PG5P5HG mixture (mix with 5% PCWS and 5% RA2) showed the best rutting performance at the three traffic levels. Yet, the rutting performance was very close for the rest of the mixtures as indicated by a relative rut depth close to one. A Hamburg Loaded-Wheel Test (LWT) was conducted in accordance to AASHTO T 324 to measure the rut depth of the five mixtures. The test was stopped when the specimen had completed 20,000 cycles or reached a 20 mm deformation. A rut depth lower than 6 mm after 20,000 cycles is desirable [28]. As shown in Fig. 4 (c), 70PG5M and 70PG5P5HG mixtures had the best rutting performance as compared to the conventional mix (70CO). Comparison of the results obtained from Pavement ME and LWT showed a good agreement in ranking for hot climate; however, a slight discrepancy was observed in cold climate. This could be related to the
Fig. 2. Master curves of asphalt mixes.
692
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
where,
Table 4 Paving materials costs.
1 2
Item
Cost
Asphalt Cement Virgin Aggregate PCWS1 & MWS2 Aggregate Rejuvenator Additive (RA1) Rejuvenator Additive (RA2)
$475/Liquid Ton $35/Ton $27/Ton $1960/Ton $2080/Ton
Jc = critical strain energy release rate (kJ/mm2); b = sample thickness (mm); a = notch depth (mm); U = strain energy to failure (N.mm); and dU/da = change of strain energy with notch depth.
PCW: Post-Consumer Waste Shingle. MWS: Manufacture Waste Shingle.
LWT test conditions, which simulate high temperature service conditions. 5.1.3. AC Bottom-up fatigue cracking Fig. 5(a and b) shows the effect of RAS on bottom-up fatigue cracking. Similar to rutting performance, 70PG5P5HG mix (mix with 5% PCWS and 5% RA2) showed the best performance, as shown in Fig. 5. The worst performer, was also the 70PG5P mix, which was the mix with 5% RAS and no RA. The mixture containing RA1 showed good fatigue performance especially at medium and high traffic levels. A semi-circular bending (SCB) test was conducted to evaluate the laboratory fracture resistance of the mixtures at intermediate temperature. In this test, semi-circular specimens with three different notch depths (25.4, 31.8 and 38.1 mm) were tested at 25 °C to evaluate the change of strain energy with notch depth (dU/da). Subsequently, the Jc value was computed by dividing dU/da by the specimen thickness:
Jc ¼
1 dU b da
ð3Þ
A typical failure criterion for the Jc is a minimum of 0.50 kJ/m2 [38]. Fig. 5c, presents the critical strain energy release rate (Jc) for the five asphalt mixtures evaluated in this study. As shown in Fig. 5c, the mixtures 70PG5P5HG and 70PG5P12CYCL were the worst performers with a Jc of 0.3 and 0.4 kJ/m2, respectively, which were lower than the minimum failure criterion. This can be explained by the increased availability of aged RAS binder in these mixes given the action of the RA. It is noted that these mixtures had the highest RBRs. Based on these results, there is an inconsistency in the prediction between Pavement ME and SCB test results since the 70PG5P5HG mixture had the best performance in Pavement ME and the worse performance in the SCB test. Based on the results presented in Fig. 5, contradicting predictions were observed between Pavement ME and the SCB test results. This is due to the stiffening effect of RAS, which is reflected in the dynamic modulus inputted in Pavement ME, and results in stiffer mixes showing better performance. To address this limitation in Pavement ME, the performance model for fatigue cracking should be modified to consider other mixture characteristics in addition to the dynamic modulus such that it can successfully reflect the fracture resistance of different classes of asphalt mixtures including the ones incorporating RAS. Based on accelerated pavement testing of different classes of asphalt mixtures, the authors have proposed a modified AASHTO model for fatigue cracking, which incorporates the critical strain energy release rate
Fig. 3. Relative terminal IRI comparison (a) cold region and (b) hot region.
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
693
Fig. 4. Relative permanent deformation comparison: (a) cold region, (b) hot region, and (c) LWT lab results.
(Jc) measured from SCB. The original fatigue model presented in Eq. (1) was modified according to the following equation [39]:
Nf ¼ k1 CC H
k2 k3 1 1 ðJ c Þk4 jE j et
ð4Þ
where, k4 = the exponent of Jc, and all the other parameters and coefficients have been defined in Eq. (1). Results showed that the modified model produced a more realistic correlation with the SCB test results and field performance [39]. 5.1.4. AC Top-down fatigue cracking Fig. 6(a and b) shows the effect of RAS on top-down fatigue cracking. Similar to bottom-up cracking performance, 70PG5P5HG mix (mix with 5% PCWS and 5% RA2) showed the best performance. The worst performer, was also the 70PG5P mix, which was the mix with 5% RAS and no RA. This trend was expected as Pavement ME deals with top-down cracking similar to bottomup fatigue cracking. 5.1.5. Service life prediction The service life of the studied cases was defined as the time at which the pavement design exceeds the terminal threshold set
for the different performance distresses. In almost all cases, the critical distress was the HMA rutting performance criterion. Fig. 7 shows the relative design life as compared to the conventional mix (70CO). As reflected in the previous discussions, mix 70PG5P5HG outperformed all the other mixes considered in the analysis. 5.2. Statistical analysis A statistical analysis was performed to assess the overall performance of the mixtures for the different combinations possible between the analyzed climate conditions and traffic levels (i.e., six iterations per mixture). An Analysis of Variance (ANOVA) was performed for each pavement distress to determine whether there was statistical difference in the mixture performance. Results of the ANOVA showed that there were significant differences between the mixture performances at a confidence level of 0.05. Therefore, a t-test was performed at a confidence level of 0.05 on all the possible combinations (i.e., 70CO vs. 70PG5P vs. 70PG5M vs. 70PG5HG) to identify the mixes that were statistically different. Table 4 shows the statistical ranking of the mixtures for the different pavement distresses. The statistical results of each grouping is ranked by using letters A, B, C, and so forth. The letter A is assigned
694
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
Fig. 5. Relative bottom-up fatigue cracking comparison: (a) cold region, (b) hot region, and (c) SCB test results.
to the mix with the highest mean followed by the letter B and so forth. Double letters (e.g., A/B, B/C) indicate that the mixture may be categorized in both groups. Results presented in Table 5 shows that the best mixture overall against roughness was 70PG5P5HG mix, which was statistically different from the rest of the mixtures. Similarly, the 70PG5P5HG mix was the best performer against rutting while mix 70CO was the worst performer in the different combinations available between the two climatic conditions and the three traffic levels evaluated in the study. A similar trend was observed for the bottom-up and top-down fatigue cracking where the best performance mixture (70PG5P5HG) was statistically different from the worst performance mixture (70PG5P). Based on the statistical results, it can be observed that the addition of 5% RA2 improved the performance of a mixture containing oxidized asphalt binder (i.e., 5% RAS) compared to the mixture with only RAS (i.e., 70PG5P and 70PG5M) and the mixture with pure virgin binder (70CO). However, the predicted performance of Pavement ME differed from the predicted SCB test results since Pavement ME does not incorporate mix resistance to cracking in the design.
5.3. Cost and energy analysis Fig. 8(a) shows the production cost of the five different mixtures. Mixtures with RAS are more economical to produce. Yet, the cost of 70PG5PCYCL was much higher due to the higher percentage of rejuvenator additive relative to the other mixtures. On the other hand, 70PG5M was the cheapest mixture to produce. The production cost was divided by the predicted service life in order to determine the mixtures that would be the most cost effective over the life cycle of the pavement. As shown in Fig. 8(b and c), 70PG5P5HG (mix with 5% PCWS and 5% RA2) had the lowest cost over the life cycle for every traffic condition and for both climatic regions. 6. Summary and conclusions This study had two main objectives: (1) to evaluate the effects of RAS (with and without recycling agents) on the predicted performance from Pavement ME; and (2) to assess the life cycle costs of pavements constructed with RAS and recycling agents.
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
695
Fig. 6. Relative top-down fatigue cracking comparison: (a) cold region and (b) hot region.
Fig. 7. Relative pavement service life: (a) cold region and (b) hot region.
An evaluation of the effects of RAS and RAs was conducted using Pavement ME for two different climatic conditions (cold and hot regions).
Based on the results of the analysis, Pavement ME predicted that the mix with 5% PCWS and 5% RA2 (70PG5P5HG) will be the best performer against roughness, HMA rutting, and fatigue
696
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
Table 5 Statistical ranking of mixtures. Mixture type
70CO 70PG5P 70PG5M 70PG5P5HG 70PG5P12CYCL
Terminal IRI
AC permanent Deformation
Bottom-up Fatigue cracking
Top-down Fatigue cracking
Mean
Rank
Mean
Rank
Mean
Rank
Mean
Rank
1.000 1.003 0.990 0.961 0.996
B A C D C
1.000 0.996 0.946 0.784 0.989
A B D E C
1.000 1.020 0.988 0.955 0.984
B A C D C/D
1.000 1.040 0.978 0.906 0.970
B A C D C/D
Fig. 8. Cost-effectiveness of mixes with RAS: (a) production cost comparison, (b) cold region, and (c) hot region.
cracking. This is due to the stiffening effect of RAS, which is reflected in the dynamic modulus inputted in the software, and results in stiffer mixes showing better performance. While one would expect RAS to improve rutting performance, the superior fatigue performance was not expected given that the binder in RAS is an air-blown asphalt binder with poor elongation and relaxation characteristics. Results of SCB test results realistically predicted that mixes with RAS and recycling agents would be the worst performers against cracking. This can be explained by the increased availability of aged RAS binder in these mixes when a
recycling agent is used. As presented in this study, the fatigue performance model in Pavement ME should be modified to consider other mixture characteristics in addition to the dynamic modulus such that it would reflect the fracture resistance of different classes of asphalt mixtures including the ones incorporating RAS. Results of the cost analysis showed that the mixtures with RAS are more economical to produce. Yet, the cost of 70PG5PCYCL was much higher due to the higher percentage of rejuvenator additive relative to the other mixtures. On the other hand, the mix with RAS and no recycling agent was the cheapest mixture to produce. The
S. Shirzad et al. / Construction and Building Materials 160 (2018) 687–697
production cost was divided by the predicted service life in order to determine the mixtures that would be the most cost effective over the life cycle of the pavement. When considering the predicted performance of the mixes, the mix with 5% PCWS and 5% RA2 had the lower cost over the life cycle for every traffic condition and for both climatic regions. References [1] L. Uzarowski, Sustainable Pavements – Making the Case for Longer Design Lives for Flexible Pavements, Annual Conference of the Transportation Association of Canada, Toronto, Ontario, 2008. [2] N.H. Tran, A. Taylor, R. Willis, Effect of Rejuvenator on Performance Properties of HMA Mixtures with High RAP and RAS Contents Report 12-05, National Center for Asphalt Technology, Auburn University, Auburn, Ala, 2012. [3] R. Karlsson, U. Isacsson, Material-related aspects of asphalt recycling—state of the art, J. Mater. Civil Eng. 18 (2006). [4] E. Brown, Preventative Maintenance of Asphalt Concrete Pavements, National Center for Asphalt Technology, Auburn University, Alabama, 1988. [5] Y.H. Huang, Pavement analysis and design, Guide for Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures: Part 1. NCHRP 1-37A Final Report, second ed., Pearson Prentice Hall, 2004. [6] A. Buss, M. Rashwan, T. Breakah, R.C. Williams, A. Kvasnak, Investigation of Warm-Mix Asphalt Using the Mechanistic-Empirical Pavement Design Guide, Mid-Continent Transportation Research Symposium, Ames, IA, 2009. [7] S.B. Cooper Jr., Characterization of HMA Mixtures Containing High Recycled Asphalt Pavement Content with Crumb Rubber Additives Thesis (MSCE), State University, 2008. [8] J.S. Daniel, G.R. Chehab, D. Ayyala, Sensitivity of RAP binder grade on performance predictions in the MEPDG, Annual Meeting and Technical Sessions CD, Association of Asphalt Paving Technologists (AAPT), Minneapolis, MN, 2009. [9] S. Diefenderfer, A. Hearon, Laboratory Evaluation of a Warm Asphalt Technology for Use in Virginia VTRC 09-R11 Final Report, Virginia Transportation Research Council, Charlottesville, VA, 2008. [10] S.W. Goh, Y. Zhanping, T.J. Van Dam, Laboratory evaluation and pavement design for warm mix asphalt, Mid-Continent Transportation Research Symposium, 2007. [11] M. Hassan, Life-cycle assessment of warm-mix asphalt: an environmental and economic perspective, Transportation Research Board 88th Annual Meeting, 2009. [12] G.W. Maupin, S.D. Diefenderfer, J.S. Gillespie, Performance and economic evaluation of virginia’s higher RAP specification, in: Transportation Research Record: Journal of the Transportation Research Board, No. 2126, 2009, pp. 142–150. [13] L.M. Mohammad, M.A. Elseifi, S.B. Cooper III, Laboratory evaluation of sulfurextended asphalt mixture, Second International Conference on Construction In Developing Countries, 2010. Cairo, Egypt. [14] S.T. Muench, Oadway construction sustainability impacts: a review of life cycle assessments, Transportation Research Board 89th Annual Meeting, 2010. [15] D. Timm, N. Tran, A. Taylor, M. Robbins, B. Powell, Evaluation of Mixture Performance and Structural Capacity of Pavements Utilizing Shell ThiopaveÒ Phase I: Mix Design, Laboratory Performance Evaluation and Structural Pavement Analysis and Design NCAT Report 09-05, National Center for Asphalt Technology, Auburn, AL, 2009. [16] National Cooperative Highway Research Program (NCHRP), Guide for Mechanistic-Empirical Design of New and Rehabilitated Structures NCHRP Report 01-37A, Transportation Research Board, Washington, DC, 2004. [17] B. Nam, H. Mahernia, Behzadan, H. Amir, Mechanical characterization of asphalt tear-off roofing shingles in Hot Mix Asphalt, Constr. Build. Mater. 50 (2014) 308–316. [18] J.R. Willis, P. Turner, Characterization of Asphalt Binder Extracted From Reclaimed Asphalt Shingles, National Center for Asphalt Technology, 2016. [19] Northeast Recycling Council, Inc., Asphalt Shingles Waste Management in the northeast Fact Sheet, 2011. http://www.nerc.oorg/documents/asphalt.pdf.
697
[20] N. Ali, J.S. Chan, A. Potyondy, R. Bushman, A. Bergen, Mechanistic evaluation of asphalt concrete mixtures containing reclaimed roofing materials, Proceedings of the 74thAnnual Meeting of Transportation Research Board, ISSN: 03611981, 1995. [21] A. Alvergue, M. Elseifi, L. Mohammad, S.B. Cooper, S. Cooper, Laboratory evaluation of asphalt mixtures with reclaimed asphalt shingle prepared using the wet process, Road Mater. Pavement Des. 15 (2014) 62–77. [22] D.E. Newcomb, M.S. Gardiner, B.M. Weikle, A. Drescher, Influence of roofing shingles and asphalt concrete mixture properties, Department of Civil and Mineral Engineering, University of Minnesota, 1993. [23] K.Y. Foo, D.I. Hanson, T.A. Lynn, Valuation of roofing shingles in hot mix asphalt, J. Mater. Civil Eng. (1999). [24] B. Sengoz, A. Topal, Use of asphalt roofing shingle waste in HMA, Constr. Build. Mater. 19 (2005) 337–346. [25] S. Tighe, N. Hanasoge, B. Eyers, R. Essex, S. Damp, Who Thought Recycled Asphalt Shingles (RAS) Needed To Be Landfilled: Why Not Build a Road?, Recycled Materials and Recycling Processes for Sustainable Infrastructure Session (Annual Conference of the Transportation Association of Canada), 2008 [26] K. Anurag, F. Xiao, S.N. Amirkhanian, Laboratory investigation of indirect tensile strength using roofing polyester waste fibers in hot mix asphalt, Constr. Build. Mater. 23 (5) (2009) 2035–2040. [27] G.W. Maupin, Investigation of the Use of Tear-Off Shingles in Asphalt Concrete, Virginia Transportation Research Council, 2010. [28] Samuel B. Cooper, Sustainable Materials for Pavement Infrastructure: Design and Performance of Asphalt Mixtures Containing Recycled Asphalt Shingles, Dissertation Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College, Baton Rouge, LA, 2015. [29] C. Robinette, J. Epps, Energy, emissions, material conservation, and prices associated with construction, rehabilitation, and material alternatives for flexible pavement, J. Transp. Res. Board (2010) 2179. [30] J.W. Arnold, B. Behnia, M.E. McGovern, B. Hill, W.G. Buttlar, H. Reis, Quantitative evaluation of low-temperature performance of sustainable asphalt pavements containing recycled asphalt shingles (RAS), Constr. Build. Mater. 58 (2014) 1–8. [31] A. Cascione, R.C. Williams, J. Yu, Performance testing of asphalt pavements with recycled asphalt shingles from multiple field trials, Constr. Build. Mater. 101 (2015) 628–642. [32] R. Ghabchi, M. Barman, D. Singh, M. Zaman, M.A. Mubaraki, Comparison of laboratory performance of asphalt mixes containing different proportions of RAS and RAP, Constr. Build. Mater. 124 (2016) 343–351. [33] Samuel B. Cooper, M. Elseifi, L. Mohammed, M. Hassan, Performance and cost effectivenes of sustainable pavement technologies in flexible pavement using the mechanistic-empirical pavement design guide, Am. Soc. Civil Eng. J. (2013). [34] S. Ishak, H. Shin, B. Sridhar, Characterization and development of truck load spectra and growth factors for current and future pavement design practices in louisiana LTRC Report 07-2P, Louisiana Transportation Research Center, Baton Rouge, LA, 2009. [35] L.M. Mohammad, A. Herath, R. Gudishala, M.D. Nazzal, M.Y. Abu-Farsakh, K. Alshibli, Development of Models to Estimate the Subgrade and Subbase Layers’ Resilient Modulus from In situ Devices Test Results for Construction Control LTRC Report 02-4B, Louisiana Transportation Research Center, Baton Rouge, LA, 2008. [36] R.L. Baus, N.R. Stires, Mechanistic-Empirical Pavement Design Guide Implementation, 2010. Report No. FHWA-SC-10-01, FHWA/SCDOT. [37] AASHTO – American Association of State Highway and Transportation Officials, Mechanistic-Empirical Pavement Design Guide – A manual of Practice, second ed., Washington D.C., 2015. [38] Z. Wu, L. Mohammad, B. Wang, M.A. Mull, Fracture resistance characterization of superpave mixture using the semi-circular bending test, J. ASTM Int. 2 (3) (2005) 1–15. [39] W. Cao, L.N. Mohammad, M.A. Elseifi, Asphalt pavement fatigue performance prediction based on semicircular bending test at intermediate temperature, Paper No. 17-05160 Presented at the 96th Transportation Research Board Annual Meeting, Washington, D.C, 2017.