Engineering Cost-Benefit Analysis of Thin Durable Asphalt Overlays

Transcription

1 Son and Al-Qadi Engineering Cost-Benefit Analysis of Thin Durable Asphalt Overlays Songsu Son 1 PhD Candidate Department of Civil and Environmental Engineering University of Illinois at Urbana-Champaign 205 N. Mathews MC250, Urbana, IL Phone: 1 (217) Fax: 1 (217) son14@illinois.edu Imad L. Al-Qadi Founder Professor of Engineering Director, Illinois Center for Transportation Department of Civil and Environmental Engineering University of Illinois at Urbana-Champaign 205 N. Mathews MC250, Urbana, IL Phone: 1 (217) Fax: 1 (217) alqadi@illinois.edu Submitted to: Transportation Research Board 93 rd Annual Meeting January 12-16, 2014 Washington, D.C. Number of Words: (12 Tables and 1 Figure) = 7421 Duplication for publication or sale is strictly prohibited without prior written permission of the Transportation Research Board 1 Corresponding author

2 Son and Al-Qadi ABSTRACT Utilizing sustainable and cost-effective pavement overlays as a rehabilitation technique have been emphasized for several years as oil prices continue their instability in the United States. Since the 1980s, significant improvements have been introduced to asphalt pavements regarding material selection, mix design, and construction technology that improve pavement performance. Some of these improvements require the use of highly modified asphalt binder and expensive high-quality aggregates to achieve desired performance. This study developed four potentially cost-effective asphalt mixtures and efficient cross-sections for asphalt wearing surfaces. The newly developed asphalt mixes incorporated locally available aggregates, special binder additives such as fibers, and/or innovative surfacing technologies. This paper presents the cost effectiveness of the newly developed asphalt mixtures while considering the improved performance. Deterministic and probabilistic life-cycle cost analysis (LCCA) was performed on the considered mixes. The performances of the considered mixes were investigated in the laboratory and on-site under actual traffic. The engineering cost-benefit analysis showed that new mixtures performed better and were more cost effective than the control mixes considered. KEY WORDS: thin asphalt overlay, life-cycle cost analysis, performance rating, and engineering cost-benefit analysis.

3 Son and Al-Qadi INTRODUCTION Using sustainable and cost-effective alternatives for asphalt wearing surfaces have been emphasized for the past several years as a result of the continuous instability of oil prices in the United States and the increase in environmental awareness. Significant improvement of pavement performances has been achieved with regard to material selection and modification, mix design, and construction technology. However, most of these improvements require highquality aggregate and expensive modified asphalt binder or special equipment for construction (1). New asphalt mixtures that offer significant cost savings were developed to improve the structural and functional conditions of the asphalt pavements, which include durability, friction, quietness, rutting resistance, cracking resistance, and moisture susceptibility resistance (2). The Bailey design method for asphalt mixtures was used to ensure proper aggregate structure of the fine dense-graded mixture gradation, thereby allowing proper compactability at a thinner layer (3). To evaluate field performance of the considered asphalt mixtures, including the control mixes, under actual traffic loading and environmental conditions, 14 pavement sections with six asphalt mixtures and various wearing surface thicknesses were constructed (4). An engineering cost-benefit analysis was performed to evaluate the cost effectiveness of the newly developed asphalt mixtures, considering their performances compared with those of the control mixes. The current cost analysis methods focus primarily on life-cycle cost analysis, which generally uses the international roughness index as a performance indicator to estimate service life of pavements. In that method, user costs are rarely taken into consideration when pavement performance is evaluated. For instance, higher surface friction is believed to increase costs because it increases fuel consumption, but the method does not take into account any benefits resulting from road safety improvements. Therefore, the engineering cost-benefit analysis in this study included overall performance of each mixture as well as total cost, including both agency costs and user costs. This study aimed to evaluate the engineering benefits of newly developed wearing surface asphalt mixtures based on performance and costs. NEWLY DEVELOPED ASPHALT MIXTURES The total cost of asphalt overlays could be reduced in various potential ways. Two reasonable approaches were considered in this study: a thinner asphalt overlay and maximum use of locally available aggregates. The layer thickness of asphalt pavements is determined by the gradation and nominal maximum aggregate size (NMAS). The minimum thickness should be four times greater than the NMAS for the stone matrix asphalt (SMA) and three times higher for the densegraded mix (5). However, to achieve target densities in the field, mixture compactability during construction should be also carefully considered. The Bailey method is a practical tool that has been successfully used to provide a better understanding of aggregate packing and to develop blend gradations that offer proper volumetric and compactability of asphalt mixtures (3). Unlike the typical coarse dense-graded mix, fine dense-graded mixes provide better compactability at a thinner layer because they have a volume of fine aggregate that exceeds the volume of voids in the coarse aggregate structure; therefore, the fine fraction carries most of the load in this mixture fraction (3). Because of the aggregate structure of fine dense-graded asphalt mixes, fine dense-graded mixes are typically easier to compact than coarse dense-graded mixes, especially if the lift thickness is near the minimum allowable layer thickness for the corresponding coarse dense-

4 Son and Al-Qadi graded mixes. Accordingly, all of the mix designs in this study were developed using the Bailey method to allow placement of the newly designed mixtures as a relatively thinner layer. Locally available aggregates were used as much as possible to ensure significant cost savings, as shown in TABLE 1. Limestone, dolomite, gravel, and steel slag are the most common aggregates in Illinois. Because it has poor durability, limestone is not permitted for use on pavements with higher traffic. Two typical Illinois surface mixtures were selected as control mixes: a 9.5-mm NMAS coarse dense-graded mixture (F-mix) and a 12.5-mm SMA. The SMA mixture generally requires more durable aggregates, highly modified asphalt binder, and cellulous fibers, which makes it more expensive than typical dense-graded mixtures. Three fine dense-graded asphalt mixes with a 9.5-mm NMAS and one SMA with a mm NMAS were developed using the Bailey method to allow placing a relatively thinner wearing course: 1. Quartzite mix contains very hard quartzite aggregate and is designed to improve pavement performance by using durable aggregate in the asphalt mix. 2. Sprinkle mix is a surface application of pre-coated coarse aggregate chips on top of a regular base mixture to improve the friction features of asphalt pavement. Costly aggregate chips with high friction are required; however, the base mixture can use less expensive aggregates. Regular dolomite and natural sand were used in the base asphalt mixture, and coarse quartzite aggregate retained on the No. 4 (4.75mm) sieve was used as sprinkle chips coated with 0.75% PG The sprinkle chips were applied at 1.56 lb/yd 2 of the spreading rate. 3. Slag/fiber mix is designed to provide good friction and high resistance to stripping and permanent deformation due to the use of steel slag. A blend of polypropylene and aramid fibers was added into this mix to improve its tensile strength; hence, allows its placement at a relatively thinner layer thickness (6, 7) mm SMA was developed to allow for a thin overlay starting at only 0.75 in, which makes it more cost effective, even though durable and expensive aggregates are required for SMA mixtures. The stone-on-stone contact of the SMA was properly verified utilizing Bailey method and through Loose Unit Weight according to AASHTO T-19. The summary of mix designs is presented in TABLE 2. TABLE 1 Aggregate Blending Percentage Mixture Type Aggregate Blending Percentage (%) Dolomite Natural Sand Quartzite Steel Slag RAP Additives Control F-Mix Mixes 12.5-mm SMA Cellulous fibers Quartzite Mix Sprinkle Mix Quartzite chips New Polypropylene Mixes Slag/Fiber Mix and aramid fibers 4.75-mm SMA Cellulous fibers Mixture Type NMAS (mm) TABLE 2 Summary of Mix Designs Gradation Binder Type Ndes Design Air Voids (%) VMA (%) AC (%)

5 Son and Al-Qadi 5 Control Mixes New Mixes F-mix 9.5 Coarse dense PG (SBS) mm SMA 12.5 SMA PG (SBS) Quartzite Mix 9.5 Fine dense PG (SBS) Sprinkle Mix 9.5 Fine dense PG (SBS) Slag/Fiber Mix 9.5 Fine dense PG (SBS) mm SMA 4.75 SMA PG (SBS) PERFORMANCE TESTS Five laboratory tests were conducted for each mixture to examine material characterizations, and five on-site tests were performed to capture the effect of pavement thickness and environmental impact under actual traffic loading. TABLE 3 is a summary of performance tests conducted in the laboratory and on-site. TABLE 3 Performance Tests Lab/Field Test Standard Output Specimen Simple Performance AASHTO TP 62 Complex modulus LMLC, PMLC Lab Moisture Indirect tensile strength AASHTO T 283 Susceptibility Tensile strength ratio LMLC, PMLC Rutting AASHTO T324 Rut depth LMLC, PMLC Durability Tex 245-F Cantabro loss LMLC, PMLC Fracture Semi-circular bending test Fracture energy LMLC, PMLC Friction AASHTO T242 Friction number On-site Noise AASHTO TP76 Onboard sound intensity On-site Field Rutting Dipstick measurement Rut depth On-site Roughness ASTM E1926 Mean roughness index On-site Texture ASTM E1845 Estimated mean texture depth On-site Laboratory Tests Lab-mixed and lab-compacted (LMLC) specimens were prepared for the new asphalt mixes, and plant-mixed and lab-compacted (PMLC) specimens were also prepared and all were used for laboratory tests. Loose asphalt materials were sampled at the plant during field construction, and the PMLC specimens were prepared after being reheated in the laboratory. The air void contents of the compacted specimens were 7% for the dense-graded mixes and 6% for the SMA; which were close to field target densities. The specimens prepared for the Cantabro Loss test were compacted at the design air void contents in accordance with the test specification. For the Sprinkle mix, the pre-coated sprinkle chips were manually spread on top of the hot mixture after flattening the surface prior to gyratory compaction. Only Cantabro Loss and wheel tracking tests were conducted on the specimens with sprinkle chips spread on the on surface. Specimens for other tests were prepared according to specifications. The sprinkle chips were applied at a rate of 1.56 lb/yd 2 which was same field application rate (8). Only Laboratory tests conducted on PMLC specimens allows better understanding of field performances. However, a longer aging time due to the reheating process might possibly have affected the test results. For the control mixes, only PMLC specimens were tested in the laboratory. The complex modulus test provides material properties over a wide range of loading frequencies and temperatures. High-temperature and low-frequency ranges usually indicate rutting potential under hot conditions and/or slow-moving loads, whereas cracking resistance may be captured at low-temperature and high-frequency ranges. Therefore, to avoid duplicating weight factors on

6 Son and Al-Qadi specific performance criteria such as rutting and fracture, complex modulus results were not considered in the overall performance score calculation. Rutting potential and cracking resistance were evaluated in separate rutting and fracture tests using a wheel tracking device and the semicircular bending test, respectively. Details on both tests may be found elsewhere (2, 4). TABLE 4 and TABLE 5 present the laboratory test results of the LMLC and PMLC specimens. Mixture Rut Depth, mm TABLE 4 Laboratory Test Results (LMLC) Fracture Cantabro Energy, Loss, % J/m 2 Indirect Tensile Strength, psi Tensile Strength Ratio, % 12 C 24 C Dry Wet Quartzite Mix Sprinkle Mix Slag/Fiber Mix mm SMA Mixture Rut Depth, mm TABLE 5 Laboratory Test Results (PMLC) Fracture Cantabro Energy, Loss, % J/m 2 Indirect Tensile Strength, psi Tensile Strength Ratio, % -12 C -24 C Dry Wet F-Mix (control) mm SMA (control) Quartzite Mix Sprinkle Mix Slag/Fiber Mix mm SMA Field Construction Fourteen field sections using the six asphalt mixtures and various thicknesses were placed over rigid pavements on IL-72 in Hoffman Estates and Barrington, Illinois, during October and November The differences between the 14 sections in the average annual daily traffic (AADT) were insignificant, although several intersections were located along the road. All tested sections have the same pavement structure; HMA overlaid jointed Portland cement concrete (JPCC). The existing pavement was repaired prior to being milled. To control reflective cracking, a leveling binder was placed over the milled surface. Hence, the variation in overlay performance due to existing pavement is insignificant. Leveling binder was placed at various thicknesses to ensure a total overlay thickness of 2 in. Because of the potential difficulties in achieving target densities for the typical control mixtures at a relatively thinner layer, the control SMA was placed at 2 in without a leveling binder, and the F-mix was placed at 1.5 in and 2 in. The target densities of all the sections were properly achieved during constructions (9). A proper aggregate packing using the Bailey method allows in situ compactability at thinner asphalt overlays. The details of field construction and layout of each section are shown in FIGURE 1.

7 Son and Al-Qadi 7 (a) (b) FIGURE 1 (a) Construction sections with different mixtures and wearing surface thicknesses, and (b) design thicknesses of leveling binder and wearing surface (after (4)). In-Place Field Testing In-place field testing was performed immediately after construction and every four months up to two years. Testing at these intervals provided results for initial field performance and short-term performance for each section. The in-place field testing included onboard sound intensity measurement, laser longitudinal texture profiling, locked-wheel friction, and walking foot inclinometer (dipstick) rut measurement. Only initial measurements immediately after construction, and one- and two-year measurements were considered in overall performance calculations. For transverse profile rut measurements, the first measurements (initial) were taken as reference, and the differences in measurements after that were reported as rut depth at each location. Therefore, four-month measurements were considered as initial rut depths. Surface cracks were visually inspected during field testing. To control reflective cracking, a leveling

8 Son and Al-Qadi binder was placed over the milled surface prior to placement of asphalt overlays in accordance with IDOT construction procedures. No significant pavement surface cracks were observed after two years of service; except section 3 exhibited reflective cracking at areas identified as segregated during construction. In-place field testing results are shown in TABLE 6.

9 Son and Al-Qadi 9 TABLE 6 On-Site Performance Test Results Section (WS/LB) Global Sound Intensity Levels, db(a) Estimated Mean Texture Depth, mm Mean Roughness Index, in./mi Initial 1 yr. 2 yr. Initial 1 yr. 2 yr. Initial 1 yr. 2 yr. 1 F-Mix (2/0) F-Mix (1.5/0.5) Quartzite Mix (1.25/0.75) Quartzite Mix (1/1) F-Mix (2/0) Slag/Fiber Mix (1/1) Slag/Fiber Mix (1.25/0.75) Sprinkle Mix (1.25/0.75) Sprinkle Mix (1/1) F-Mix (2/0) mm SMA (0.75/1.25) mm SMA (1/1) mm SMA (1.25/0.75) mm SMA (2/0) Friction Number Friction Number Section (WS/LB) (Smooth Tire) (Treaded Tire) Rut Depth, mm Initial 1 yr. 2 yr. Initial 1 yr. 2 yr. 4 mo. 1 yr. 2 yr. 1 F-Mix (2/0) F-Mix (1.5/0.5) Quartzite Mix (1.25/0.75) Quartzite Mix (1/1) F-Mix (2/0) Slag/Fiber Mix (1/1) Slag/Fiber Mix (1.25/0.75) Sprinkle Mix (1.25/0.75) Sprinkle Mix (1/1) F-Mix (2/0) mm SMA (0.75/1.25) mm SMA (1/1) mm SMA (1.25/0.75) mm SMA (2/0)

10 Son and Al-Qadi 10 Overall Performance Rating Calculation It is important to select an appropriate method to quantify the overall performance of each mix to allow for reasonable comparison. To evaluate each mix s performance in the laboratory and field, statistical analysis was performed using Fisher s least significant difference (LSD) method. Fisher s LSD test was performed to compare the mean value of each group and to rank them using Statistical Analysis System (SAS) software. The test was performed with two-way analysis of variance (ANOVA) at a significance level of 0.05 for each mixture s property and performance based on laboratory and field test results. The test results were ranked using the letters from A to E for laboratory test results and A to J for field test results; more groups with different thicknesses were considered in the field testing results. The letter was determined by statistical difference of the mean from others the letter A represents the best performing mixture followed by the other letters in alphabetic order. For the rankings of noise and mean roughness index, the letter A represents the least noise and roughness, respectively. Fisher s LSD test was performed on laboratory performance test results for the LMLC and PMLC specimens and field performance test results for initial, one- and two-year measurements, except for the rut measurement, which had an initial value of at four months. The alphabetic ranks for each mixture with LMLC and PMLC specimens and the field performance test results at zero, one, and two years for each asphalt mixture and thickness are shown in TABLE 7 (a) to (e). Once alphabetic ranks were determined from Fisher s LSD test for each mixture and thickness, those ranks were converted to numbers to calculate the overall performance numerically. The scale of numerical scores was determined from 0 to 10 (worst to best) with two intervals for each alphabetic rank of laboratory test and one interval for field test results starting with a score 10 for rank A. A double letter, such as A/B, in the Fisher s LSD results indicates that the difference in the means was not statistically significant and that the mixture s ranking could fall in either group. The overall performance score was calculated using Equation 1. This equation allows the decision maker to apply a weight factor ( ai or a j ) for a specific performance value when calculating the overall performance score. For example, rut resistance could be weighted more heavily in high-temperature locations, while fracture resistance could be emphasized more by applying a higher weight factor in cold locations. In addition, if the field test results are considered more important, a greater weight factor ( ) could be applied. In this study, the weight factors were assumed to be the same for all performance and property values for both the laboratory and field tests. The average value of the test results was used when the test was conducted at more than one condition, such as fracture test ( 12 C and 24 C), indirect tensile strength (dry and wet), and friction (treaded tire and smooth tire). n m L L F F a i R a j R i j 1 i 1 j 1 Equation 1 Overall Performance Score n m L F ai a j i 1 j 1 where, L F a, a : Weight factor of test i and j for lab performance and field performance, respectively i j

11 Son and Al-Qadi 11 R, R L i F j : Performance rating of test i and j for lab performance and field performance, respectively, : Weight factor for lab performance and field performance, respectively n, m : Number of tests performed in the lab and the field, respectively TABLE 8 shows the overall performance score for each section. In general, the newly developed asphalt mixes have greater indirect tensile strength and less noise than the control mixes. In general, the control mixes resulted in lower performance scores compared to the new mixtures. The 4.75-mm SMA section with a 1-in wearing surface provided the highest overall performance score.

12 Son and Al-Qadi 12 TABLE 7 Performance Ranking for Laboratory Test Results: (a) LMLC and (b) PMLC; and Field Test Results: (c) Initial, (d) 1 Year, and (e) 2 Year (a) Lab-Mixed and Lab-Compacted (LMLC) Fracture Indirect Tensile Mix Type Rut Complex Modulus Durability Resistance Tensile Strength Strength Resistance 12 C 24 C Dry Wet Ratio 10 C, 25 Hz 54 C, 0.1 Hz Quartzite Mix A B/C B A/B A A B A A/B Sprinkle Mix B C A/B A/B A B B A A Slag/Fiber Mix A/B A/B A/B B A A/B B A B 4.75-mm SMA C A A A B C A A B (b) Plant-Mixed and Lab-Compacted (PMLC) Fracture Indirect Tensile Mix Type Rut Complex Modulus Durability Resistance Tensile Strength Strength Resistance 12 C 24 C Dry Wet Ratio 10 C, 25 Hz 54 C, 0.1 Hz F-Mix (Control) A B B B D E A A B SMA (Control) A C A A C/D D/E A A/B A Quartzite Mix A B B B B/C B/C A B B Sprinkle Mix A B B/C B A A A A/B B Slag/Fiber Mix A A/B C B B A/B A A/B B 4.75-mm SMA A A B B B/C C/D A A/B B

13 Son and Al-Qadi 13 (c) Section (WS/LB)* Initial Noise Friction (Tread) Friction (Smooth) Rut Resistance MRI EMTD F-Mix (1.5/0.5) B/C F C B C/D C F-Mix (2/0) C/D E/F C B C/D C 12.5-mm SMA (2/0) E D/E C B F A Quartzite Mix (1/1) C/D C B B A D Quartzite Mix (1.25/0.75) D B B B/C A/B D Sprinkle Mix (1/1) C/D A A A C/D C/D Sprinkle Mix (1.25/0.75) C/D B B A/B C/D D Slag/Fiber Mix (1/1) A/B C C/D B A/B D Slag/Fiber Mix (1.25/0.75) C/D C C B C/D D 4.75-mm SMA (0.75/1.25) A C A/B B D/E B 4.75-mm SMA (1/1) A C/D B C D/E B/C 4.75-mm SMA (1.25/0.75) A/B G D A/B E C * Indicates wearing surface thickness/leveling binder thickness. Section (WS/LB) 1 Year Noise Friction (Tread) Friction (Smooth) Rut Resistance MRI EMTD F-Mix (1.5/0.5) F E C B A/B B F-Mix (2/0) F D C/D B B B/C 12.5-mm SMA (2/0) G C A B C A Quartzite Mix (1/1) C E G B B E Quartzite Mix (1.25/0.75) D A/B E/F B/C A/B F/G Sprinkle Mix (1/1) E/F A D/E A A C/D Sprinkle Mix (1.25/0.75) E D/E F A/B B D/E Slag/Fiber Mix (1/1) D B G B A/B F Slag/Fiber Mix (1.25/0.75) D E G B B G 4.75-mm SMA (0.75/1.25) B B A/B B B C/D 4.75-mm SMA (1/1) A A/B A/B C B C 4.75-mm SMA (1.25/0.75) C C B A/B B D (d)

14 Son and Al-Qadi 14 (e) Section (WS/LB) 2 Years Noise Friction (Tread) Friction (Smooth) Rut Resistance MRI EMTD F-Mix (1.5/0.5) F A B/C A/B A/B B F-Mix (2/0) E/F B C/D B/C A/B B/C 12.5-mm SMA (2/0) G A A A/B B A Quartzite Mix (1/1) C/D C D/E B A/B D Quartzite Mix (1.25/0.75) D/E A F/G B/C A/B D Sprinkle Mix (1/1) E/F C E/F A/B A B Sprinkle Mix (1.25/0.75) E/F C/D G A A/B B/C Slag/Fiber Mix (1/1) C/D C H A A D Slag/Fiber Mix (1.25/0.75) C/D D G/H B A/B D 4.75-mm SMA (0.75/1.25) A/B B B/C C A/B B/C 4.75-mm SMA (1/1) A A/B B B A/B B/C 4.75-mm SMA (1.25/0.75) B/C B E/F A/B A/B C TABLE 8 Overall Performance Score Section (WS/LB) Rut Fracture Durability Resistance Resistance IDT TSR Noise Friction MRI EMTD Overall F-Mix (1.5/0.5) F-Mix (2/0) mm SMA (2/0) Quartzite Mix (1/1) Quartzite Mix (1.25/0.75) Sprinkle Mix (1/1) Sprinkle Mix (1.25/0.75) Slag/Fiber Mix (1/1) Slag/Fiber Mix (1.25/0.75) mm SMA (0.75/1.25) mm SMA (1/1) mm SMA (1.25/0.75)

15 Son and Al-Qadi 15 LIFE-CYCLE COST ANALYSIS Life-cycle cost analysis (LCCA) is a technique for evaluating the overall economic efficiency of a project or product. LCCA incorporates initial costs and discounted future agency, use, and other relevant costs over the life of pavements such as maintenance, rehabilitation, restoring, resurfacing, and reconstruction costs. LCCA was performed on six asphalt mixtures and various wearing surface thicknesses. RealCost software, developed by the Federal Highway Administration (FHWA), was used for the deterministic and probabilistic analyses of LCCA (10). Agency cost includes material cost and construction cost, which were obtained from the company involved in the field construction. RealCost software provided agency cost and user cost for each pavement section considering the various asphalt mixtures at different wearing surface thicknesses. Agency Costs The agency costs include all costs related to raw material, production, installation, maintenance, and replacement. For this study, the costs for maintenance and replacement for all study mixtures were assumed to be equal because no life cycles for new mixtures were available at the time. TABLE 9 shows agency costs, which include material cost, production cost, and construction cost per ton of asphalt mixture. The control SMA shows higher agency costs because of using binder in this mix and relatively high steel slag content. The additional screening of typical stockpiles for quartzite and dolomite aggregates on a 4.75-mm (No. 4) sieve to achieve proper aggregate gradation resulted in a higher material cost of the 4.75-mm SMA. The control F-mix had higher recycled material content: 10% reclaimed asphalt pavement (RAP), and 35.7% and 84% of steel slag for the F-mix and 12.5-mm SMA, respectively. An additional $2 per ton of asphalt mix was added to the construction cost of the sprinkle mix to account for the cost of the spreading equipment and trucking of sprinkle chips. TABLE 9 (a) shows the agency cost for each considered asphalt mixture. Material quantity per lane-mile for various layer thicknesses was calculated with respect to the layer thickness and required density. Total agency cost (including material, production, and construction costs) for each section was then calculated by multiplying the material quantity by the cost per lane-mile as shown in TABLE 9 (b). The test section with the control 12.5-mm SMA and a 2-in wearing surface was the most expensive section in this study, while the quartzite mix section with a 1-in wearing surface was the least expensive. To simplify the calculation of unit total cost per lane-mile, the analysis period was set at 1 year; the section length was set at 1 mile, with 0.25 mile of work zone transition before and after the section. Both deterministic and probabilistic analyses were performed in this project, and the probabilistic analysis followed a normal distribution.

16 Son and Al-Qadi 16 TABLE 9 Agency Cost for Each (a) Mix, and (b) Section (a) Mix Type Cost, $/ton Materials + Production Construction Total F-Mix (Control) mm SMA (Control) Quartzite Mix Sprinkle Mix Slag/Fiber Mix mm SMA Leveling Binder (b) Cost, $/Lane-Mile Section (WS/LB)* Materials + Production Materials + Production + Construction Quartzite Mix (1/1) 42,722 59,074 Quartzite Mix (1.25/0.75) 42,791 59,407 Sprinkle Mix (1/1) 43,115 60,296 Sprinkle Mix (1.25/0.75) 43,197 60,813 Slag/Fiber Mix (1/1) 44,369 61,076 Slag/Fiber Mix (1.25/0.75) 44,851 61, mm SMA (0.75/1.25) 49,262 65,288 F-Mix (1.5/0.5) 47,967 66, mm SMA (1/1) 51,534 67,804 F-Mix (2/0) 49,808 68, mm SMA (1.25/0.75) 53,807 70, mm SMA (2/0) 76,597 97,587 * Indicates wearing surface thickness/leveling binder thickness. User Cost The user cost is primarily based on the time delay by travelers as a result of pavement rehabilitation. Its major components are related to the time delay cause by the work zone speed limit and any related increased vehicle operation cost. The construction time for each test section was assumed to be the same, from 7 a.m. to 3 p.m. Likewise, the same work zone speed limit of 25 mph was applied to each section, based on the Illinois Vehicle Code (65 ILC S 5/11-604). The normal speed limit for the test sections was 35 mph; the 10 mph reduction in speed resulted from the pavement work. The user cost for a test section was computed using FHWA s RealCost LCCA software. The hourly traffic distribution followed the national average as provided in the software. The value of user time per vehicle class from 1996 was converted to reflect 2011 costs using the allitems consumer price index (CPI) (10). As shown in TABLE 10 (a), the CPI was in 1996 and in 2011, according to the Bureau of Labor Statistics, which comes up with the escalation factor of between

17 Son and Al-Qadi and 2011 (calculated by dividing the 2011 CPI by the 1996 CPI). The value of user time in 2011 was then obtained using that escalation factor, as shown in TABLE 10 (b). TABLE 10 (a) Consumer Price Index for 1996 and 2011, and (b) User Time Values for 2011 (a) Consumer Price Index (CPI) All-Items Component, All-Items Component, Escalation Factor (b) Value of User Time $/hr Year 2011 Vehicle Class Minimum Most Likely Maximum Passenger Vehicles Single-Unit Trucks Combination-Unit Trucks Deterministic Analysis A deterministic cost analysis was performed using a fixed cost without allowing any change or variation of input parameters. The user cost for 2-in F-mix and 12.5-mm SMA sections was almost half that of other sections because the 2-in F-mix and 12.5-mm SMA sections were placed with a single lift, whereas the others had two lifts the leveling binder and the wearing surface. Therefore, the sections with two lifts required traffic control twice, which doubled the user cost. After the user costs were added, the control 12.5-mm SMA with a 2-in wearing surface was the most expensive section, at $101,210 per lane-mile of total cost, and the quartzite section with a 1-in wearing surface was the least expensive section, almost 35% less costly than the control 12.5-mm SMA. In general, the new mixtures were less expensive than the control mixes except for the 4.75-mm SMA sections with a 1-in and a 1.25-in wearing surface. The relatively high total cost of the 4.75-mm SMA was caused primarily by additional screening of the typical stockpile aggregate to obtain proper gradation and its relatively high asphalt binder content of 7.3%. Although steel slag was used almost 84% of the aggregate by weight in the control mm SMA its highly modified asphalt binder contributed to its relatively high total cost, as shown in TABLE 11 (a). Probabilistic Analysis Probabilistic cost analysis considers the risk and uncertainty of input parameters such as traffic status change and material cost change. This helps decision makers minimize the risk of uncertainty of different variables. TABLE 11 (b) presents the probabilistic cost for each section, calculated using RealCost software. The quartzite mix section provided the lowest total cost among the evaluated sections. The uncertainty has more of an impact on user cost, with its higher standard deviation, than on the agency cost. The minimum and maximum total costs represent the best- and worst-case scenarios that can occur for the tasks analyzed.

18 Son and Al-Qadi 18 TABLE 11 Total Cost: (a) Deterministic, and (b) Probabilistic (a) Section (WS/LB*) Agency Cost, $1000 User Cost, $1000 Total Cost, $1000 Quartzite Mix (1/1) Quartzite Mix (1.25/0.75) Sprinkle Mix (1/1) Sprinkle Mix (1.25/0.75) Slag/Fiber Mix (1/1) Slag/Fiber Mix (1.25/0.75) F-Mix (2/0) mm SMA (0.75/1.25) F-Mix (1.5/0.5) mm SMA (1/1) mm SMA (1.25/0.75) mm SMA (2/0) (b) Agency Cost, User Cost, Total Cost, $1000 $1000 $1000 Section (WS/LB) Standard Standard Standard Mean Mean Mean Min. Max. Deviation Deviation Deviation Quartzite Mix (1/1) Quartzite Mix (1.25/0.75) Sprinkle Mix (1/1) Sprinkle Mix (1.25/0.75) Slag/Fiber Mix (1/1) Slag/Fiber Mix (1.25/0.75) mm SMA (0.75/1.25) F-Mix (1.5/0.5) F-Mix (2/0) mm SMA (1/1) mm SMA (1.25/0.75) mm SMA (2/0) * Indicates wearing surface thickness/leveling binder thickness. ENGINEERING COST-BENEFIT ANALYSIS The engineering cost-benefit analysis for an asphalt overlay is a performance-based cost analysis that considers both total cost (including agency and user costs) and overall performance of asphalt overlay. It provides a key input for cost-effectiveness evaluation. The engineering cost per unit performance in this study was calculated by dividing total cost by the overall performance scores, as shown in TABLE 12. Total cost per unit performance represents the initial cost of providing one unit overall performance for each mixture and section. The quartzite mix with a 1-in wearing surface was the most cost-effective section. All of the sections with new mixtures had a much lower cost per unit performance than the control mixtures, although the control mixes contained a significant amount of recycled materials such as steel slag and RAP. No recycled materials were used in the

19 Son and Al-Qadi 19 new mixes for this project. However, if recycled materials are considered for use in the new mixes, the total cost is expected to be decreased unless the addition of recycled materials results in a performance reduction. The layer thickness does not significantly affect a mixture s performance for this study. Therefore, the thinner layer was more cost effective than the thicker wearing surface in terms of cost per overall unit performance. TABLE 12 Cost Per Unit Performance Section (WS/LB)* Agency Cost, Overall Cost Per Unit Performance, $/Lane-Mile Performance Score $/Lane-Mile/Performance Quartzite Mix (1/1) 59, ,231 Sprinkle Mix (1/1) 60, ,256 Quartzite Mix (1.25/0.75) 59, ,334 Slag/Fiber Mix (1/1) 61, , mm SMA (0.75/1.25) 65, ,444 Sprinkle Mix (1.25/0.75) 60, , mm SMA (1/1) 67, ,696 Slag/Fiber Mix (1.25/0.75) 61, , mm SMA (1.25/0.75) 70, ,342 F-Mix (1.5/0.5) 66, ,508 F-Mix (2/0) 68, , mm SMA (2/0) 97, ,495 * Indicates wearing surface thickness/leveling binder thickness.

20 Son and Al-Qadi SUMMARY AND FINDINGS This study evaluated pavement performances of newly developed mixes in terms of rut resistance, moisture susceptibility, durability, fracture resistance, friction, noise, and ride quality. The study assessed the new mixes ability to improve pavement performance while reducing construction and use costs. An engineering cost-benefit analysis was performed to evaluate the cost effectiveness of the new asphalt overlays at various thicknesses. The results of the laboratory and field performance tests showed that the new mixtures performed better and were more cost effective than the control mixes. The following points summarize the conclusions of this study. Newly developed mixes including fine dense-graded and 4.75-mm SMA were properly placed and compacted at relatively thinner layers, with thicknesses ranging from 0.75 in to 1.25 in. The Bailey method, used for the mix design, provided a proper aggregate structure for the new mixtures to allow in situ compactability and reduction of layer thickness. The quartzite mix and sprinkle mix provided excellent cost savings with promising pavement performance when using a relatively thinner wearing surface layer. Given that Illinois imports quartzite, the use of sprinkle mix becomes more efficient. The overall performance ratings showed that the newly developed mixes provided significantly better overall performance than the control mixtures, especially for noise level and indirect tensile strength. Even though the control mixtures contained significant amounts of recycled materials, total costs of the new mixes were less than the control mixes, except for the sections of the 4.75-mm SMA with a 1-in and a 1.25-in wearing surface. Additional costs were needed for screening the aggregates on a No. 4 (4.75-mm) sieve to achieve target aggregate gradations. This cost may be eliminated as the use of the mix becomes common. The quartzite mix and sprinkle mix with a 1-in-thick wearing surface layer provided the lowest cost per unit performance. The sprinkle treatment is expected to have a more sustainable impact because most aggregates used are local. All of the newly developed mixes are more cost effective than the control mixes. The control 12.5-mm SMA had the highest cost per unit performance. The use of recycled materials such as recycled asphalt shingles and RAP in the newly developed mixes is expected to reduce the total cost while maintaining overall performance. ACKNOWLEDGMENTS This publication is based on the results of ICT-R27-42, Development of a Thin, Quiet, Long- Lasting, and High Friction Surface Layer for Economical Use in Illinois. ICT-R27-42 was conducted in cooperation with the Illinois Center for Transportation; the Illinois Department of Transportation, Division of Highways; and the U.S. Department of Transportation, Federal Highway Administration. The contents of this paper reflect the view of the authors, who are responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Illinois Center for Transportation, the Illinois Department of Transportation, or the Federal Highway Administration. This paper does not constitute a standard,

21 Son and Al-Qadi specification, or regulation. Special thanks are to Tom Zehr and David Lippert of IDOT for all the help they provided as well as other technical review panel members including Abdul Dahhan, LaDonna Rowden, Sheila Beshears, Hal Wakefield, Bill Pine, and John Lavalee. Special thanks are also to colleagues who helped in the project including Jeff Kern, James Meister, and ICT students. Part of the field measurements was conducted by Applied Research Associates (ARA), and their help is appreciated. REFERENCES 1. Estakhri, C. K., and J. W. Button. Evaluation of Ultrathin Friction Course. Transportation Research Record, No. 1454, 1994, pp Al-Qadi, I. L., S. Son, and S. H. Carpenter. Development of an Economical, Thin, Quiet, Long-Lasting, and High Friction Surface Layer Volume 1: Mix Design and Lab Performance Testing. FHWA-ICT , Illinois Center for Transportation, Urbana, IL, Vavrik, W. R., Huber, G., Pine, W. J., Carpenter, S. H., and Bailey, R. Bailey Method for Gradation Selection in Hot-Mix Asphalt Mixture Design. Transportation Research E-Circular: Journal of the Transportation Research Board, Vol. E-C044, 2002, pp Al-Qadi, I. L., S. Son, and T. Zehr. Development of an Economical, Thin, Quiet, Long-Lasting, and High Friction Surface Layer Volume 2: Field Construction, Field Testing, and Engineering Benefit Analysis. FHWA-ICT , Illinois Center for Transportation, Urbana, IL, Brown, E. R., M. R. Hainin, L. A. Cooley Jr., C. Abadie, M. Heitzman, B. Ruth, J. Scherocman, K. Tam, M. Dunning, M. Marasteanu, R. Gribbin, M. Taylor, and F. Rodriguez. Determining Minimum Lift Thickness for Hot Mix Asphalt (HMA) Mixtures. In 2005 Meeting of the Association of Asphalt Paving Technologists, March 7, March 9, Association of Asphalt Paving Technologist, Long Beach, CA, United states, 2005, pp Bueno, M., J. Luong, U. Vinuela, F. Teran, and S. E. Paje. Pavement Temperature Influence on Close Proximity Tire/Road Noise. Applied Acoustics, Vol. 72, No. 11, 2011, pp Lee, S. J., J. P. Rust, H. Hamouda, Y. R. Kim, and R. H. Borden. Fatigue Cracking Resistance of Fiber-Reinforced Asphalt Concrete. Textile Research Journal, Vol. 75, No. 2, 2005, pp Son, S., I. Al-Qadi, D. Lippert, and T. Zehr. Innovative Sprinkle Treatment for Thin Durable Asphalt Overlays. Transportation Research Record: Journal of the Transportation Research Board, Vol. 2366, No. -1, 2013, pp Leng, Z., I. Al-Qadi L, P. Shangguan, and S. Son. Field Application of Ground-Penetrating Radar for Measurement of Asphalt Mixture Density: Case Study of Illinois Route 72 Overlay. Transportation Research Record: Journal of the Transportation Research Board, No. 2304, 2012, pp. pp

22 Son and Al-Qadi Walls III, J., and M. R. Smith. Life-Cycle Cost Analysis in Pavement Design Interim Technical Bulletin. FHWA-SA , Federal Highway Administration, Washington, DC,