Long-Term Performance of Flexible Pavements Constructed on Recycled Base Layers

Abstract

There has been an increased interest in using reclaimed asphalt pavement (RAP) aggregates as a base course in Wisconsin for the offered economic and environmental benefits. Laboratory studies showed that RAP can have resilient modulus values higher than typical natural aggregates, and can also have higher durability, especially in freeze–thaw cycles. However, it is also recognized that RAP exhibits temperature sensitivity and larger permanent deformations than natural aggregates. How these characteristics manifest themselves in the northern U.S. climates can only be assessed by long-term observation of field performance. Wisconsin Department of Transportation (WisDOT) has been using RAP as a base course for over 30 years. The qualitative assessment of WisDOT roads constructed with RAP showed they are performing adequately. However, this impression has not yet been verified quantitatively. This paper presents a quantitative assessment and comparisons for field performance of pavement test sections constructed with RAP and natural crushed aggregate (CA) base course layers in Wisconsin. The performance was evaluated using pavement distress surveys of roadways containing RAP and CA, falling weight deflectometer (FWD), dynamic cone penetrometer (DCP) tests, and analysis of WisDOT’s available data. Based on the field evaluation, the performance of the hot mix asphalt (HMA) pavements with RAP was found adequate and comparable to those with CA base course layers. A recommendation was given that WisDOT shall continue the practice of using RAP in base course layers of HMA pavements, given proper quality checks for materials, and proper quality control/quality assurance measures being applied during construction.

Keywords

flexible pavements long-term performance reclaimed asphalt pavements natural crushed aggregates pavement distresses field evaluation base courses recycled materials

In flexible pavement systems, the base course layer acts to distribute traffic loads to the underlying subbase and subgrade layers, as well as to facilitate drainage. The base course must also provide support to the wearing surface to prevent tensile fatigue cracking. Base course aggregate must have adequate permeability, durability, and angularity. Pertinent properties of unbound aggregate are characterized by parameters such as resilient modulus, saturated hydraulic conductivity, strength such as California bearing ratio (CBR), maximum dry density, and optimum water content. These parameters are critical for a mechanistic-empirical pavement design method.

In 2020, 1.46 billion tons of crushed stone valued at more than $17.8 billion were produced by 3,440 quarries operating in the U.S. ( 1 ). These figures raise issues related to sustainability, as quarries gradually become depleted and environmental regulations become more stringent. With high demand for construction aggregate and an increasing public desire to manage waste materials in a responsible manner, there has been increasing interest in the prospect of utilizing by-products and reclaimed material for pavement construction purposes. For base courses in flexible pavement systems, the use of reclaimed asphalt pavements (RAP) and recycled concrete aggregates (RCA) has been the subject of increased research in recent years ( 2 ).

Tremendous amounts of RAP are generated from road resurfacing projects. The 2020 annual report of the National American Pavement Association (NAPA) estimated that 97 million tons of RAP are reclaimed by their industry for future use ( 3 ). Over 100 million tons of RAP are milled annually in the United States ( 4 ). The amount of RAP produced often exceeds the amount that can feasibly be reused for hot mix asphalt (HMA) mixes ( 5 ). Excess RAP generated on site can be utilized near the site as a base course material, thus reducing material transportation costs. RAP is pulverized and processed on site to reach the desired gradation.

RAP possesses several characteristics that make it an attractive choice for use in unbound base course material. Several studies have shown that the resilient modulus of RAP is higher than that of virgin aggregate (6, 7). Kim and Labuz ( 8 ) found that pavements with various percentages of RAP in the base course performed similarly with respect to strength and stiffness compared with pavements with 100% virgin aggregate base courses.

RAP tends to have a high fine content because of the crushing and milling process. Fine RAP particles tend to be coated with hydrophobic asphalt, which reduces the ability to hold excess moisture (9, 10). Nokkaew et al. ( 11 ) found the saturated hydraulic conductivity values for RAP range between 3.7 × 10⁻⁵ m/s and 3.7 × 10⁻⁴ m/s, when compacted to 95% of maximum dry density. As asphalt is a viscoelastic binding agent, RAP tends to show deformation over time under sustained stresses (creep), and numerous studies have shown that pavements with RAP base courses are susceptible to rutting (4, 12). Dong and Huang ( 4 ) found that RAP showed larger permanent deformation than limestone aggregate and gravel following repeated triaxial load testing. Researchers have also noted the relative weakness of RAP aggregates. Bennert et al. ( 6 ) found that natural aggregates had higher shear strength than pure RAP mixes. Thakur et al. ( 5 ) found a decrease in CBR values in RAP with increasing binder content and decreasing fines content. Bleakley and Cosentino ( 12 ) found that the limerock bearing ratio (LBR), a variation of the CBR test, for RAP fell short of acceptable limits; the researchers found that strength issues can be mitigated by using RAP blended with natural aggregates or by means of chemical stabilization.

Several researchers conducted laboratory testing to evaluate RAP and mixtures of RAP with virgin aggregates. Arshad and Ahmed ( 13 ) evaluated the properties of RAP blends containing 50% and 75% RAP with virgin aggregates and RCA. Tests to measure resilient modulus (M_R) and constrained modulus (M_C) were conducted. Samples containing 75% RAP/25% aggregate showed a significant increase in M_R values, especially under greater bulk stresses. M_C values, on the other hand, decreased with higher RAP proportions. An increase in the percentage of RAP led to an increase in the accumulated strains under cyclic loading. Bradshaw et al. ( 14 ) evaluated the resilient moduli of RAP and virgin aggregate blends. Subbase samples from Route 165 in Rhode Island were tested. Route 165 was reconstructed during the 1980s from an unbound mixture of cold recycled RAP and virgin aggregates blended off-site, and it was reconstructed in 2013 using full-depth reclamation (FDR). The samples were compacted at optimum moisture content and 95% max dry density for the resilient modulus test. The M_R of the untreated FDR RAP blends was higher than that of the cold recycled RAP because of the greater RAP content. Compared with the cold recycled RAP blend, these samples exhibited greater shear softening and permanent strains. Similar studies were conducted by Kim and Labuz ( 8 ), who reported that RAP samples showed a permanent deformation at least two times greater than 100% CAs but declared that base materials containing different RAP proportions performed similarly to 100% aggregates in regard to resilient modulus and strength when properly compacted.

According to Stolle et al. ( 15 ), shear strength is slightly reduced, and deformation is increased when RAP is mixed with natural aggregates, as determined by triaxial tests. Lukanen and Kruse ( 16 ) recommend a maximum RAP content of 50% in base layers, claiming that RAP proportions greater than that lead to reduced strength and increase rutting.

NCHRP Synthesis 598 studied the feasibility of using RAP as a construction material in unbound layers ( 17 ). Laboratory tests conducted on blends of RAP and other virgin aggregates indicated that based on toughness, stiffness, and strength characteristics based on monotonic triaxial testing, RAP was identified as a suitable construction material in unbound aggregate layers for high-volume roads in nonfreezing conditions. Note that documented freeze–thaw behavior of RAP is inconsistent in the literature. Bozyurt ( 18 ) found the summary resilient modulus (SRM) of RAP decreased at a relatively rapid rate after five freeze–thaw cycles, with a decreasing rate after subsequent cycles. Attia and Abdelrahman ( 9 ), however, reported an increase in SRM after freeze–thaw cycling.

Further, many states limit the proportion of reclaimed material that can be used in pavement base courses. In Wisconsin, Section 301 of WisDOT Standard Specifications for Highway and Structure Construction provides specifications for base course aggregate. Extensive testing of pavement with a reclaimed base course has been conducted at the MnROAD facility. The Minnesota Department of Transportation (MnDOT) boasts over 30 years’ experience in flexible pavement construction using RAP and suggests that adherence to best practices will produce a pavement that can outperform pavements with natural aggregate base courses ( 19 ).

Objective and Scope

The intended outcome of this study is to conduct surveys to collect and analyze pavement distress for HMA roadways constructed in Wisconsin using RAP as a base course aggregate and to compare with similar roadways constructed with natural aggregates to verify the performance of roadways constructed with recycled base aggregates. For any negative attributes to the use of RAP, the presented results from field evaluation provide methods to determine whether there are any techniques that can be utilized to produce satisfactory results using recycled aggregates.

Description of Field Evaluation Methods

The following section provides the methodology to select a comprehensive testing matrix for roadways in Wisconsin that utilize RAP and natural coarse aggregates in base layers, and to describe the testing program and methods used to evaluate the pavement sections.

Selection of Pavement Test Sites

The criteria used for the selection of sites considered three aspects: (1) geographical variation in Wisconsin, (2) base course layers that used virgin crushed stone aggregates (CA) and RAP, and (3) HMA pavement type. The selected pavement sites are HMA pavements with CA and RAP base courses that were constructed in or earlier than 2009. In total, six pavement sections with CA base courses, and six pavement sections with RAP base courses were identified and used to compare performance in this paper. Three highway projects that contain segments with CA and RAP base courses were studied in more detail to compare and contrast performance. These three pavements are presented in this paper as three case studies and are described in detail.

The six project sites with CA base course layers are: State Trunk Highway (STH) 22/54 (Waupaca), STH 22 (Shawano), STH 33 (St Joseph), STH 77 (Webb Lake), STH 25 (Maxville), and STH 59 (Edgerton). The pavement cross sections for the pavements with CA base course layers are shown in Figure 1. The six project sites with RAP base course layers are: STH 22 (Shawano), STH 70 (Minocqua), STH 96 (Lark-Shirley), STH 59 (Edgerton), STH 25 (Maxville), and STH 77 (Webb Lake). The pavement cross sections for the pavements with RAP base course layers are shown in Figure 2. The layer thicknesses were obtained from the WisDOT project plans and measurement by the research team during pavement coring. The three project sites that have both CA and RAP base course segments and that are discussed in detail in this paper as case studies are STH 59 (Edgerton), STH 25 (Maxville), and STH 77 (Webb Lake).

Figure 1.

Typical cross sections for the investigated HMA pavements with CA base courses: (a) STH 22/54 (Waupaca), (b) STH 22 (Shawano), (c) STH 33 (St Joseph), (d) STH 77 (Webb Lake), (e) STH 25 (Maxville), and (f) STH 59 (Edgerton).

Figure 2.

Typical cross sections for the investigated HMA pavements with RAP base courses: (a) STH 22 (Shawano), (b) STH 70 (Minocqua), (c) STH 96 (Lark-Shirley), (d) STH 59 (Edgerton), and (e) STH 25 (Maxville).

Field Evaluation Methods

The testing program consisted of falling weight deflectometer (FWD), visual distress surveys, pavement surface profile measurements, and dynamic cone penetration (DCP). Note that only some results from FWD and DCP testing are shown in this paper for brevity. The full data was presented elsewhere ( 20 ).

The following presents a summary of the field tests conducted at the investigated pavement sections:

Falling Weight Deflectometer (FWD): The FWD test was conducted using a KUAB^® FWD according to ASTM D4694: Standard Test Method for Deflections with a Falling-Weight-Type Impulse Load Device. The FWD was used with three different load drops of 22.2, 40.0, and 53.4 kN (5, 9, and 12 kips, respectively). Seven geophones were used to record pavement surface deflection, located at the center of the loading plate and at 30.5, 61, 91, 122, 152, 183 cm (12, 24, 36, 48, 60, and 72 in., respectively) behind the loading plate. Pavement surface, air temperatures, and GPS coordinates were acquired at each test point. The FWD test data was analyzed using the pavement layer moduli back-calculation software developed by ERI, Inc. The analysis was conducted using pavement layer thicknesses obtained from the WisDOT project plans and measurement by the research team during pavement coring. Pavement deflections were first normalized to the 40 kN (9,000 lb) load and then adjusted for temperature variations.

Visual and Automated Pavement Surface Distress Surveys: Visual surveys were conducted to identify and quantify the various types of pavement surface distress exhibited at the investigated pavements and to obtain data needed to evaluate pavement performance as a pavement condition index (PCI). Each distress survey was conducted for one 161 m (528 ft) section at each pavement site. The section was selected to be representative of the overall pavement condition. It should be noted that the WisDOT Pavement Data Unit conducts automated pavement surface distress surveys as part of pavement management of the state/national highway network. The collected data is compiled in the Pavement Information/Inventory Files (PIF) database where the performance indicators such as the PCI and the International Roughness Index (IRI) are calculated for the length of the fourth 160 m (0.1 of a mile) for each highway segment. The research team accessed the PIF database and analyzed the data corresponding to pavement sections investigated in this study.

At the investigated pavement sites, surface distresses were visually identified, quantified, and recorded. Pavement distress types, extent, and levels of severity were identified and quantified according to the FHWA distress identification manual. Obtained via personal communication with WisDOT Pavement Data Unit, the following limits were considered for the various distresses and performance indicators to assess the performance :

PCI: very good to excellent (85–100); good (70–84); fair (55–69); poor (0–54)

Longitudinal and transverse cracking: low (average crack width < 9.5 mm or a filled crack of any width); medium (average crack width > 9.5 mm or crack surrounded by secondary cracking); high (average crack width > 76 mm or crack is severely spalled/broken)

Rutting: low (6.4 mm–12.7 mm); medium (12.7 mm–25.4 mm); high (larger than or equal to 25.4 mm)

Alligator/fatigue cracking: low (a singular longitudinal crack in the wheel path or multiple wavy longitudinal cracks in the wheel path running parallel to each other with no, or little in the way of, interconnecting cracks); medium (interconnecting cracks forming an “alligator” pattern of cracks with no dislodgements, occurring anywhere on the roadway); high (interconnected cracks causing dislodgement and rocking pieces, occurring anywhere on the roadway)

Pavement Surface Profile Measurements: Pavement surface profile measurements were conducted using the CS8800 Walking Profiler System provided by Surface Systems & Instruments, Inc. The profile measurements were conducted on the inside wheel path, center of the lane, and outside wheel path for a length of 183 m (600 ft) at each investigated pavement test section. The system is equipped with GPS system and software that allows for real-time display of measured profiles.

Dynamic Cone Penetration Test: The field testing program included aggregate base course layer and subgrade testing using the DCP. Equipment with a single-mass hammer was used, and the DCP was driven into the aggregate base layer (through the HMA hole). The test was conducted according to the standard test procedure described by ASTM D6951: Standard Test Method for Use of the DCP in Shallow Pavement Applications. DCP tests were conducted at the wheel path and lane center HMA holes in which the cone was driven through the whole aggregate base course layer and into the subgrade. The CBR and resilient modulus (M_R) of the base layers were estimated from the DCP penetration rate (PR) in millimeters per blow using the Webster et al. (21, 22) and the Powell et al. ( 23 ) equations for CBR and M_R, respectively—see Equations 1 and 2.

CBR (%) = \frac{292}{{PR}^{1.12}}

(1)

M_{R} = 17.58 CB R^{0.64}

(2)

Further, the research team retrieved the base materials from the selected pavement sites after performing the previously described field tests. Base material samples with a volume of approximately one to two 19 L (5 Gal) buckets (depending on the site condition) were collected from these sites.

Evaluated Case Studies

In this section three case studies of highway pavements in the state of Wisconsin that contain pavement sections with RAP and natural virgin aggregates are assessed comprehensively. These case studies are: (1) STH 59 west of Edgerton, Rock County, (2) STH 25 near Maxville, Buffalo County, and (3) STH 77 near Webb Lake, Burnett County. Figures 1 and 2 show the pavement structures for these pavements.

Case 1: STH 59 West of Edgerton, Rock County, Wisconsin

Project Description

This project consists of two sections: (1) pavement section I with a RAP base layer that was reconstructed in 2009 with an original CA base layer before 2009, and (2) pavement section II with a CA base layer that was milled and resurfaced in 2009. The total length of the project is 11.7 mi (18.8 km) of which approximately 1 mi has a CA base layer, and the remainder has a RAP base layer.

Field work was conducted on both pavement sections in 2009, 2010, 2011, 2017, and 2018 that included visual distress surveys, FWD, and DCP testing. The results of the field tests as well as the analyses conducted on pavement data obtained from WisDOT are presented here. Figure 3 presents pictures of the pavement surface condition at both test sections before and after reconstruction.

Figure 3.

STH 59 pavement sections with RAP and CA base layers before reconstruction (2009) and after reconstruction (2011 and 2018): (a) pavement section I with CA base layer in 2009 before mill and relay, (b) pavement section II with CA base layer in 2009 before mill and overlay, (c) mill and relay in 2009, (d) transverse cracking in 2011 in pavement section II with CA base layer section, and (e) longitudinal cracking in RAP base layer section in 2018.

Project Evaluation—Field Testing

The results from the analysis of FWD tests on STH 59 pavement section I are presented in Figure 4. The FWD testing was conducted on a 46 m (150 ft) long section with measurements taken at every 3 m (10 ft) interval on both eastbound (EB) and westbound (WB) lanes, at the right and left wheel paths. The testing configuration provided detailed data pertaining to pavement surface deflection and structural capacity. Figure 4 indicates relatively high pavement surface deflection (with the maximum FWD center deflection, D₀ ranging approximately from 254 to 610 µm (10 to 24 mils), which was expected because of the pavement surface deterioration. Figure 4a shows higher pavement surface deflections at the pavement edges than at the centerline, and also shows high variability in the back-calculated moduli for the HMA, base and subbase layers (E_HMA, E_base, E_subgrade, respectively).

Figure 4.

Results of FWD test conducted in 2009 on STH 59 pavement section I with crushed aggregate base, just before milling and relay of the HMA layer: (a) adjusted normalized deflection under FWD plate, D₀, in mils, (b) back-calculated HMA layer modulus, E_HMA, in ksi, (c) back-calculated base layer modulus, E_base, in ksi, and (d) back-calculated subgrade modulus, E_subgrade, in ksi.

Figure 5 shows a comparison between the pavement structural capacities of section I with a CA base layer (before reconstruction in 2009) and the same section with a RAP base layer after reconstruction in 2010. The contour maps of the effective structural number (SN_eff) presented in Figure 5 (under the similar color scale) indicated higher structural capacity 1 year after reconstruction of the pavement with a RAP base layer (SN_eff ranging from 4.1 to 6.0) compared with the same pavement section with a CA base layer (SN_eff ranging from 5.6 to 7.4). It should be noted that the pavement section before reconstruction had a CA base layer of unknown age and the testing on the same section with a RAP base layer was conducted 1 year after reconstruction. Figure 6 presents box-whisker plots of the pavement effective structural capacity variation with time for the CA and RAP base layer sections at STH 59 from 2009 to 2017. An examination of the figure demonstrates that, on average, the pavement section with a RAP base layer exhibited higher structural capacity (and lower FWD deflections) compared with the same section before reconstruction.

Figure 5.

Comparison of structural capacity of pavement section I on STH 59 before reconstruction (CA base layer in 2009) and 1 year after reconstruction (RAP base layer—mill and relay 2009): (a) structural capacity of pavement section I in 2009 when the base was CA and (b) structural capacity of section I in 2010 when the base was reconstructed into RAP.

Figure 6.

Pavement structural capacity for STH 59 with a CA base layer before reconstruction to a RAP base layer after reconstruction.

The results of the long-term pavement performance analysis for STH 59 with RAP and CA base layers indicated that the pavement section with CA exhibited lower rutting at the left wheel path, had no longitudinal cracking, and had lower transverse cracking than the pavement segments with a RAP base layer. For all segments (CA and RAP), rut depth in pavement surveys conducted in 2009, 2011, 2013, 2015, and 2017 was less than 5 mm (0.2 in.). In general, the performances of the pavement segments with CA and RAP base layers were comparable with IRI values less than 1.42 m/km (90 in./mile), except for one pavement segment with a RAP base that exhibited high IRI values. The change in pavement performance indicators with time is depicted in Figure 7. The PCI variation for the pavement segments is depicted in Figure 7, a and b , showing good performance level for all segments with PCI greater than 70. The change in the average PCI for each pavement section (CA and RAP base layer sections) is depicted in Figure 7c with a comparably good performance, noticing that the CA base has a slightly higher average PCI. The average increase in IRI with time is comparable for both pavement sections, while the pavement section with RAP exhibited slightly higher rutting but less than an average of 4.6 mm (0.18 in.) as presented in Figure 7d. The analysis of STH 59 pavement performance data shows that, on average, the pavement segment with a CA base layer slightly outperformed the pavement segments with a RAP base layer; however, both pavement sections can be given a good performance rating.

Figure 7.

Comparison of the performance of the HMA pavement segments of STH 59 constructed on CA and RAP base layers for: (a) average PCI for all segments over the pavement life, (b) overall average of PCI segments with CA and RAP base layers over the pavement life, (c) ride quality—IRI (in./mi) versus age, and (d) rutting versus pavement age.

Case 2: STH 25 Near Maxville, Buffalo County, Wisconsin

Project Description

The pavement test section at STH 25 near Maxville in Buffalo County consists of 18.8 km (11.7 mi) section with a RAP base course layer and one 730 m (2,394 ft) long section with a crushed stone aggregate base built in 2004. The pavement test sections on STH 25 provide an excellent example comparison between the performance of a virgin crushed aggregate and a recycled base course since both CA and RAP sections were constructed and the sections have similar climatic condition, age, subgrade, and traffic levels. The CA base section has slightly higher HMA thickness (127 mm [5 in.] versus 114 mm [4.5 in.]) but a thinner base thickness (305 mm [12 in.] versus 381 mm [15 in.]) with dense-graded aggregates and a nominal maximum aggregate size of 31.8 mm (1.25 in.). Laboratory particle size distribution tests on the CA and RAP base layer materials showed that the STH 25 CA material had higher gravel size and fines fractions and lower sand size fraction than the STH 25 RAP material. The CA materials showed an average mass loss of 18.8% in a Micro-Deval abrasion test and an average absorption of 2.91%. The average absorption for the RAP material was 2.3%. A Micro-Deval test was not conducted on the RAP base material.

Project Evaluation—Field Testing

Field evaluation conducted on both sections included DCP and FWD. The strength characterization with DCP showed that the CA base layer had lower penetration rates and higher back-calculated CBR and strength values than the corresponding values for the RAP section. The average predicted CBR (by DCP test) for the CA layer was 90.1% with a layer modulus of 314 MPa (45.5 kips per square inch [ksi]) with corresponding values of 75.3% and 278 MPa (40.3 ksi) for the RAP layer. On the other hand, the pavement section with the RAP base layer had a lower average FWD center deflection, D₀ of 221 µm (8.7 mils) and higher average effective structural number, SN_eff of 6.3 when compared with average D₀ of 241 µm (9.5 mils) and SN_eff of 5.6 for the CA section. These results are consistent with the back-calculated FWD base layer moduli for both sections, where the average E_base was 548 MPa (79.5 ksi) for the RAP base and 303 MPa (44 ksi) for the CA base layer. In PCI calculations, the test section (0.1 mi section surveyed visually) with the CA base showed significant fatigue cracking, resulting in an average PCI of 29 compared with 74 for the RAP test section (0.1 mi section surveyed visually). The same trend was observed when calculating the IRI from the walking profiles on both test sections with an average IRI of 1.65 m/km (104.6 in./mi) for the CA base layer section and 0.99 m/km (63 in./mi) for the RAP base layer section.

The long-term pavement performance was evaluated by WisDOT Pavement Data Unit by conducting pavement surveys in 2009, 2011, 2013, 2015, and 2017 for both CA and RAP pavement sections. The data was analyzed by the research team. Data includes pavement surface condition (PCI), ride quality/smoothness of ride (IRI), rutting, and cracking (fatigue, longitudinal, transverse, edge, and block cracking).

Figure 8 presents the performance of pavement segments for rutting and ride quality with age, obtained as sequences from WisDOT’s PIF database, showing the CA and RAP base layer segments. A chip seal treatment was applied in 2017 at pavement age of 13 years, but it neither improved the ride quality nor reduced the rutting of the RAP and CA base layer segments at the time of measurement, except for a slight improvement in the rutting performance in the measurements taken at the left wheel path. Similarly, Figure 9 presents the fatigue, edge, longitudinal, and transverse cracking records for the different segments. The cracking in the sections with the RAP base are, on average, similar to those in the CA sections. Overall, an inspection of the pavement performance indicators in Figures 8 and 9 show that the pavement segment with a CA base layer (29140) had a performance comparable with the RAP base layer segments (29130, 29150, and 29160) located nearby, but outperformed the other segments located at the beginning of the pavement project. Note that the pavement segment with a CA base layer outperformed the pavement segments with a RAP base layer on average; however, the difference is not very significant. It should be noted that there is only one pavement segment with a CA base layer that is approximately 730 m (2,394 ft) in length compared with 18.8 km (11.7 mi) of segments with a RAP base course. Note also that based on visual observation along the total length of the project, the research team believes that the pavement with the RAP base showed good performance comparable with the CA base, if not better.

Figure 8.

Comparison of rutting and ride quality (IRI) performance for STH 25 segments with CA and RAP base layers: (a) rutting in right wheel path, (b) rutting in left wheel path, (c) IRI in right wheel path, and (d) IRI in left wheel path.

Figure 9.

Comparison of the cracking and PCI performance for STH 25 segments with CA and RAP base layers: (a) alligator (fatigue) cracking, (b) edge cracking, (c) longitudinal cracking, (d) transverse cracking, (e) pavement condition index versus time, and (f) pavement condition index along the segments.

Case 3: STH 77 near Webb Lake, Burnett County, Wisconsin

Project Description

The flexible pavement segments of STH 77 near Webb Lake consist of two parts: a 14.8 km (9.2 mi) segment with 114 mm (4.5 in.) HMA surface layer constructed on a 152 mm (6 in.) RAP base layer followed, on the east direction, by a 7.4 km (4.6 mi) pavement segment with a 127 mm (5 in.) thick HMA surface layer constructed on 254 mm (10 in.) dense-graded CA base layer. The project consisted of pavement reconstruction of CA base layer segments in the year 2011 followed by the reconstruction of the pavement with a RAP base layer in the year 2012.

Project Evaluation—Field Testing

The long-term pavement performance indicators for rutting and ride quality are presented for pavement segments with both CA and RAP base layers in Figure 10. Generally, both pavement types performed well since they are relatively newly constructed and showed insignificant rutting and good ride quality. However, as shown in Figure 11, the pavement segments with a RAP base layer developed a fair amount of longitudinal and transverse cracking compared with the pavement segments constructed on a CA base layer. Despite having higher cracking levels, most of the cracks were of low severity. In fact, the PCI values for the pavement segments with a RAP base layer still fell within the “good” range of PCI, and all segments with RAP bases had a PCI exceeding 80 (Figure 11, c and d ).

Figure 10.

Comparison of rutting and ride quality (IRI) performance for STH 77 segments with CA and RAP base layers: (a) rutting in right wheel path, (b) rutting in left wheel path, (c) IRI in right wheel path, and (d) IRI in left wheel path.

Figure 11.

Comparison of the cracking performance and PCI for STH 77 segments with CA and RAP base layers: (a) total longitudinal cracks (low severity), (b) total transverse cracks (low severity), (c) pavement condition index along the segments, and (d) pavement condition index versus time.

Comparisons of All Investigated Pavement Sections

The results of the analysis of the PIF database pertaining to all investigated pavement sections with CA and RAP base layers are presented here. An inspection of the data in Figure 12 indicates that the PCI variation with time did not show a clear trend among the pavement test sections with CA and RAP base layers. However, a pavement section with a CA base layer exhibited the lowest PCI rating 20 years after construction. The ride quality data demonstrated that the base sections with RAP base layers generally performed better compared with the pavement sections with CA base layers. In respect of the average rutting, some of the pavement sections with RAP base materials exhibited the highest rut depth. Alligator (fatigue) cracking was observed in higher quantities in pavement sections with RAP base materials compared with the sections with CA base layers. Transverse cracking occurred more often in pavement sections with RAP base layers at an older age compared with pavement sections with RCA base materials. Longitudinal cracking was more visible in the pavement sections with RCA base layers at a younger age compared with pavement sections with CA and RAP base layers.

Figure 12.

Average performance indicators for AC and RAP base layers: (a) average PCI, (b) average IRI, (c) average rut depth, (d) average fatigue cracking, (e) average transverse cracking, and (f) average longitudinal cracking, plotted for the total length of the project versus pavement age.

To better visualize the trends in performance between RAP and CA base layer pavement sections, Figure 13 present the variations of the averages of pavement performance indicators (PCI, IRI, rutting, and cracking) with HMA pavement age for all investigated pavements with CA and RAP base layers. The results of long-term pavement performance analyses conducted here using the WisDOT PIF database demonstrate that the performance of the HMA pavements with RAP is comparable with the performance of the HMA pavements with CA base course layers (commonly used in WisDOT projects).

Figure 13.

Average performance indicators for all investigated pavement projects with AC and RAP base layers: (a) average PCI, (b) average IRI, (c) average rutting (depth), (d) average alligator (fatigue) cracking, (e) average transverse cracking, and (f) average longitudinal cracking.

Summary and Conclusions

The Wisconsin Department of Transportation (WisDOT) has been using RAP as a base course for over 30 years. The qualitative assessment of WisDOT roads constructed with RAP base layers with HMA surfaces concluded that they are performing adequately. This paper intended to provide quantitative evaluation of the use of RAP as a base layer in HMA pavements. The study used several techniques to collect data via field testing programs on pavement sections from HMA pavements with natural CA and RAP base layers. Such techniques included FWD, pavement surface profile measurements using a walking profiler, visual pavement distress surveys, DCP tests, and analysis of pavement performance data from WisDOT’s PIF database for the investigated pavement sections.

Based on the results of evaluating field case studies with CA and RAP base layers, the following conclusions could be made:

The pavements with CA base layers exhibited the highest deflection and the least variability.

The pavement test sections with CA base layers had the lowest average effective structural number (SN_eff) values. The pavement test sections with RAP base layers exhibited intermediate SN_eff values.

In general, the back-calculated base layer moduli (E_base) for all investigated pavement tests indicate that the RAP base layers had higher moduli than CA base layers. Subgrade moduli (E_Subgrade) values for the investigated pavement sections were comparable and fell within a close range of values.

Based on the visual distress survey of the investigated pavement sections, fatigue cracking is the most commonly observed surface distress associated with CA and RAP base layers. Note that the whole pavement structure will contribute to fatigue cracking, and the observed cracks may be more related to the HMA layer than they are to the base materials.

Based on the results of this study, the performance of the HMA pavements with RAP is satisfactory and adequate, and comparable with the performance of the HMA pavements with CA base course layers. Based on the results of field investigations, it is recommended that WisDOT continues the practice of using RAP in base course layers of HMA pavements, but some measures should be considered to ensure adequate performance. This includes blending proportions of RAP and CA—for example, 50/50 could be used, which is in line with the recommendations and observations of other researchers. Further, some construction requirements can be implemented as quality assurance and control measures. These include but are not limited to: density checks, DCP profiling, and utilizing Light Weight Deflectometer (LWD) or Geogauge^® tests for stiffness checks on top of the constructed base course layers.

Footnotes

Acknowledgements

This publication is based on the results of WHRP 0092-17-01 Project: “Evaluation of Recycled Base Aggregates” ( 20 ). WHRP 0092-17-01 was conducted in cooperation with the Wisconsin Department of Transportation, and the U.S. Department of Transportation, Federal Highway Administration. The authors would like to acknowledge and thank Mr. Daniel Reid and Ms Tracy Petersen of WisDOT for their help, support, and effort during field testing. The input and guidance of WHRP Geotechnical Oversight Committee members and WisDOT engineers Mr Andrew Zimmer, Mr Robert Arndorfer, and Mr Jeffrey Horsfall, is greatly appreciated. The authors would also like to thank Mr Joseph Allaby, Mr Michael Wolf, and Mr Andrew Schilling of WisDOT Pavement Data Unit for their help and support.

Author Contributions

The authors confirm contributions to the paper as follows: study conception and design: Hani Titi, Habib Tabatabai; data collection: Hani Titi, Issam Qamhia; analysis and interpretation of results: Hani Titi, Habib Tabatabai, Jessie Ramirez, Issam Qamhia; draft manuscript preparation: Hani Titi, Issam Qamhia. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Wisconsin Department of Transportation Project number: WHRP 0092-17-01.

ORCID iDs

Hani Titi

Issam I. A. Qamhia

Habib Tabatabai

Data Accessibility Statement

The data that supports the findings of this study is available from the corresponding author, Hani Titi, on reasonable request.

The contents of this paper reflect the view of the authors, who are responsible for the facts and the accuracy of the data presented here. The contents do not necessarily reflect the official views or policies of the WisDOT. This paper does not constitute a standard, specification, or regulation.

References

United States Geological Survey. Crushed Stone Statistics and Information: 2021 Annual Report. USGS, 2021. https://pubs.usgs.gov/periodicals/mcs2021/mcs2021-stone-crushed.pdf.

Gabr

Cameron

Properties of Recycled Concrete Aggregate for Unbound Pavement Construction. ASCE Journal of Materials in Civil Engineering, Vol. 24, No. 6, 2012, pp. 754–764. https://doi.org/10.1061/(ASCE)MT.1943-5533.0000447.

National Asphalt Pavement Association. 2020 Annual Report. NAPA, Greenbelt, MD, 2020. https://www.asphaltpavement.org/?option=com_content&view=article& id=201&Itemid=495.

Dong

Huang

Laboratory Evaluation on Resilient Modulus and Rate Dependencies of RAP Used as Unbound Base Material. ASCE Journal of Materials in Civil Engineering, Vol. 26, No. 2, 2014, pp. 379–383. https://doi.org/10.1061/(ASCE)MT.1943-5533.

Thakur

S. C.

Han

Chong

W. K.

Parsons

R. L.

Laboratory Evaluation of Physical and Mechanical Properties of Recycled Asphalt Pavement as a Base Course Material. Proc., GeoShanghai International Conference, Shanghai, China, June 3–5, 2010.

Bennert

Papp

Jr. Maher

Gucunski

Utilization of Construction and Demolition Debris under Traffic-Type Loading in Base and Subbase Applications. Transportation Research Record: Journal of the Transportation Research Board, 2000. 1714: 33–39.

Bejarano

Evaluation of Recycled Asphalt Concrete Materials as Aggregate Base. Technical Memorandum prepared for the California Department of Transportation, District 2 Materials Branch. UCPRC-TM-2001-04.? Pavement Research Center, Institute of Transportation Studies, University of California, Berkeley. 2001, p. 23.

Kim

Labuz

Resilient Modulus and Strength of Base Course with Recycled Bituminous Material. Research Report No. 2007-5. Minnesota Department of Transportation, Minneapolis, 2007, p. 270.

Attia

Abdelrahman

Sensitivity of Untreated Reclaimed Asphalt Pavement to Moisture, Density, and Freeze Thaw. ASCE Journal of Materials in Civil Engineering, Vol. 22, No. 12, 2010, pp. 1260–1269.

10.

Edil

Tinjum

Benson

Recycled Unbound Materials. Research Report No. 2012-35. Minnesota Department of Transportation, Minneapolis, 2012, p. 340.

11.

Nokkaew

Tinjum

Benson

Hydraulic Properties of Recycled Asphalt Pavement and Recycled Concrete Aggregate. In ASCE GeoCongress 2012: State of the Art and Practice in Geotechnical Engineering ( Hryciw

R. D.

Athanasopoulos-Zekkos

Yesiller

, eds.), American Society of Civil Engineers, Reston, VA, 2012, pp. 1476–1485. https://doi.org/10.1061/9780784412121.152.

12.

Bleakley

Cosentino

Improving Properties of Reclaimed Asphalt Pavement for Roadway Base Applications through Blending and Chemical Stabilization. Transportation Research Record: Journal of the Transportation Research Board, 2013. 2335: 20–28.

13.

Arshad

Ahmed

Potential Use of Reclaimed Asphalt Pavement and Recycled Concrete Aggregate in Base/Subbase Layers of Flexible Pavements. Construction and Building Materials, Vol. 151, 2017, pp. 83–97.

14.

Bradshaw

Costa

Giampa

Resilient Moduli of Reclaimed Asphalt Pavement Aggregate Subbase Blends. ASCE Journal of Materials in Civil Engineering, Vol. 28, No. 5, 2016, p. 04015206. https://doi.org/10.1061/(ASCE)MT.1943-5533.0001508.

15.

Stolle

Guo

Emery

Mechanical Properties of Reclaimed Asphalt Pavement - Natural Aggregate Blends for Granular Base. Canadian Journal of Civil Engineering, Vol. 41, No. 6, 2014, pp. 493–499. https://doi.org/10.1139/cjce-2013-0009.

16.

Lukanen

E. O.

Kruse

C. G.

Optimization of Recycled Asphalt Concrete and Base Materials for Bases of New Asphalt Concrete Pavements. Report No. SD97-03. South Dakota Department of Transportation, Pierre, 2000.

17.

Saeed

. Performance-Related Tests of Recycled Aggregates for Use in Unbound Pavement Layers. National Cooperative Highway Research Program (NCHRP) Report No. 598. Transportation Research Board, National Research Council, Washington, DC, 2008.

18.

Bozyurt

Behavior of Recycled Asphalt Pavement and Recycled Concrete Aggregate as Unbound Road Base. Master’s thesis. Department of Civil and Environmental Engineering, University of Wisconsin-Madison, 2011.

19.

Johnson

Clyne

Recycled Asphalt Pavement: MnROAD Study of Fractionated RAP. Minnesota Local Research Board (LRRB) Investigation 864 and, MPR Project No. 6-(022). Minnesota Department of Transportation, St. Paul, 2012.

20.

Titi

H. H.

Tabatabai

Ramirez

Sooman

Evaluation of Recycled Base Aggregates. Report No. WHRP 0092-17-01. University of Wisconsin—Milwaukee, 2019.

21.

Webster

S. L.

Grau

R. H.

Williams

T. P.

Description and Application of Dual Mass Dynamic Cone Penetrometer. Instruction Report No. GL-92-3. US Army Engineer Research and Development Center, Waterways Experiment Station, Vicksburg, MS, 1992, p. 50.

22.

Webster

S. L.

Brown

R. W.

Porter

J. R.

Force Projection Site Evaluation Using the Electronic Cone Penetrometer (ECP) and the Dynamic Cone Penetrometer (DCP). Technical Report No. GL-94-17. US Army Engineer Research and Development Center, Waterways Experiment Station, Vicksburg, MS, 1992, p. 172.

23.

Powell

W. D.

Potter

J. F.

Mayhew

H. C.

Nunn

M. E.

The Structural Design of Bituminous Roads. TRRL Report No. LR 1132. Transport and Road Research Laboratory, Wokingham, Berkshire, 1984, p. 62.