Annotated Literature Review for NCHRP Report 640 (2009)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

237 and performance before establishing a minimum value for number of contact points needed to ensure stone-on stone contact; but this method seems promising. The authors refer to the use of image analysis for optimizing the gradations for different aggregates with different properties, through the determination of stone-on-stone contacts and establishing both a minimum and maximum number of contacts per cubic meter. 1.60.8 Structural Design No information is provided on structural design. 1.60.9 Limitations No information is provided on limitations of use. 1.61 Bennert, T., F. Fee, E. Sheehy, A. Jumikis and R. Sauber. “Comparison of Thin-Lift HMA Surface Course Mixes in New Jersey” TRB 2005 Annual Meeting CD-ROM. Transportation Research Board. National Research Council. Washington, D.C. 2005. 1.61.1 General Bennert et al. provides a comparison of performance and cost-effectiveness of different types of pavement surfaces used in New Jersey. They describe results of tests conducted to evaluate performance related to noise, skid resistance and roughness, and also provide cost information. The different types of surfaces include modified open-graded friction course (MOGFC), Asphalt Rubber OGFC (AROGFC), 12.5 mm nominal maximum aggregate size Superpave Hot Mix Asphalt (HMA), microsurfacing (specifically, Novachip), Stone Matrix Asphalt (SMA), and unfinished, diamond ground and transverse tined Portland Cement Concrete (PCC). The authors provide objectives of using different surfaces, descriptions of the different pavement structures for these surfaces, types of field tests, description of winter maintenance activities for the OGFC, and comparison of results and costs. No particular surface showed the best performance in all tests. For example the transverse tined PCC did not show the high performance in roughness or noise reduction but did show the best performance in skid resistance. Bennert et al. indicate that the overall performance and cost comparison showed that the OGFC provides the most cost effective surface, followed by the 12.5 mm HMA and the diamond ground PCC. They note that the second best performers, the HMA and the PCC cost significantly more than the OGFC because of its (OGFCs) relatively thin lift thickness and also because the diamond grinding can be conducted only on intact PCC surfaces.

238 1.61.2 Benefits of Permeable Asphalt Mixtures No information on specific benefits of PFCs is given in the paper. However, in discussing the objectives of using relatively thin lifts in New Jersey pavements, the authors refer to certain desirable qualities, that are mostly present in PFCs. For example, Bennert et al. mentions that the New Jersey Department of Transportation (NJDOT) wishes to use a mix that can reduce noise, improve skid resistance, provide a smooth ride and at the same time can be placed in relatively thin lifts (25 mm). The fact that these pavements are on very busy routes rule out the possibility of using cold paving materials (which need time for curing), and NJDOT would like to use a mix that can maximize the coverage area, that is, in per unit of time. These mixes are viewed as “maintenance” mixes and the authors note that one of the desirable qualities is good durability, which would help NJDOT avoid frequent disruption of traffic. 1.61.3 Materials and Design The mix design information on the open-graded friction courses (and other surface courses) is shown in Tables 114 through 116. Note that the authors provide mix design information on several mixes used in several projects, only three of which have been described in detail (for construction). Table 114: Mix Design Information Mix/Project/Design Year Property AROGFC, I-1 95, Jackson/ 1992 MOGFC, US 24, Union/ 1999 MOGFC-2, I 195, Hamilton Square/ 2001 Novachip, I 195, Hamilton Square/2001 Gradation, percent passing Sieve Size, mm 19 12.5 9.5 4.75 2.36 1.18 0.6 0.3 0.15 0.075 100 94 39 10 7 5 4 4 3.0 100 86 57 13 9 6 5 4 3 2.8 100 91 74 18 7 4 3 3 2 2.0 100 78 27 22 16 13 10 7 5.9 Design Method 25 blow Marshall 50 gyration SGC 50 Gyration SGC 50 Gyration SGC Design air voids, % 18 21.1 23.0 12.0 Additive/Modifier 15% GTR (#80), Wet Process PG 76-22 (PMA) + 0.4 % Fiber PG 76-22 (PMA) + 0.4 % Fiber PG 76-22 (PMA) + Modified Emulsion Tack Asphalt Binder Content, % 6.60 6.80 6.0 5.00 Note: The projects in which these four mixes were used have been described in detail

239 Table 115: Mix Design Information Continued Mix/Project/Design Year Property AROGFC/US 9, N/1993 MOGFC, I 78 W/2002 9.5 mm SMA, I 78 E/2002 Micro-S, Type 3, US 202 S, NJ 29/2002 Gradation, percent passing Sieve Size, mm 19 12.5 9.5 4.75 2.36 1.18 0.6 0.3 0.15 0.075 100 88 36 12 8 6 5 4 3.0 100 91 59 15 6 4 3 3 2 1.8 100 93 45 25 18 16 15 13 11.6 100 96 72 52 36 22 15 10.7 Design Method 25 blow Marshall 50 gyration SGC 50 Gyration SGC --- Design air voids, % 16.5 21.0 4.0 --- Additive/Modifier 15% GTR (#40), Wet Process PG 76-22 (PMA) + 0.3 % Fiber PG 76-22 (PMA) + 0.4 % Fiber Latex modified Asphalt Binder Content, % 6.50 6.20 6.90 6.00 Table 116: Mix Design Information Continued Mix/Project/Design Year Property 12.5 mm SP, US 22W/2000 12.5 mm SP, I 78 E/2003 12.5 MM SMA, US 1 N/S/1993 Gradation, percent passing Sieve Size, mm 19 12.5 9.5 4.75 2.36 1.18 0.6 0.3 0.15 0.075 100 88 72 44 30 22 17 11 7 4.9 100 98 86 53 33 24 18 13 8 4.9 100 85 67 38 27 22 19 16 12 9.7 Design Method 125 gyration SGC 125 gyration SGC 25 blow Marshall Design air voids, % 4.0 4.2 3.3 Additive/Modifier PG 76-22 (PMA) PG 76-22 (PMA) AC 20 Asphalt Binder Content, % 4.80 4.90 6.00 1.61.4 Construction Practices Bennert et al. provides description of the different layers, some construction details and costs for each of the sections evaluated. The descriptions are summarized in Table 117.

240 Table 117: Description of Pavement Structure and Construction Project/Type Description I-195 – Jackson Township/four lane interstate US 24 – Union/ four lane interstate I-195 – Hamilton Square/four lane interstate Location 30 miles west of Trenton. 15 miles west of NY City. 5 miles west of Trenton. Original pavement 6.5 inches of HMA, constructed in 1972 9 inches of PCC, constructed in 1974 Not given Traffic, ESAL for 20 years 27 million 70 million 27.8 million Intermediate course 3 inches of NJ I-2 HMA base --- 4-5 inches of 12.5 mm & 19 mm Superpave Cause of rehabilitation Not Given Reduce noise Enhance the structural number Construction date Sept. 1992 Paving: Fall, 1999 Reflection cracks noted on surface is Spring, 2000; Saw and seal done; Novachip – Oct. 2000 / MOGFC-2 –May 2001 Overlay design Mill 2 inches and fill 4 inches 1 ¼ inches + a heavy application of hot (PG64- 22) tack Mill 2 inches / replace 5-6 inches Surface course 1 inch of Asphalt Rubber Open Graded Friction Course (AR-OGFC). A #80 mesh Ground Tire Rubber (GTR) was used in the “wet” process. Modified Open Graded Friction Course #1 gradation WB – ½” Novachip®; EB ¾” MOGFC-2 Cost of surface course $7.30/yd2 $4.60/yd2 Cost of Novachip = $2.62/yd2; Cost of OGFC-2 = $1.76/yd2 1.61.5 Maintenance Practices Bennert et al. mention that the cracks in the PCC pavement on the US-24 Union section had caused reflection cracks in the MOGFC between Fall of 1999 and Spring of 2000, and these cracks were subsequently sawed and sealed. Regarding winter maintenance activities, the authors mention that the NJDOT uses rock salt for deicing, whereas the New Jersey Garden State Parkway (NJGSP) uses liquid magnesium chloride. The authors note that that the NJDOT finds the OGFC mixes to be more difficult to maintain ice-free than the adjacent dense-graded mixes, even though more frequent application of rock salts are made for these mixes. The NJGSP, on the other hand, had good success with the magnesium chloride, although the OGFC needs twice the amount of application as the dense-graded mix. The NJGSP continually monitors forecasts of temperature and measures surface temperatures, and pretreats OGFC surfaces with liquid magnesium chloride to avoid icing. If the magnesium chloride is applied after the OGFC is frozen then it washes off the surface. However, by pretreating, NJGSP has found that the OGFC surfaces are manageable and can be plowed the same as the dense graded mixes.

241 1.61.6 Rehabilitation Practices No information on rehabilitation practices is provided in this paper 1.61.7 Performance Bennert et al. provides the results of performance of different types of surfaces, including OGFCs, in terms of noise reduction, skid resistance and ride quality. The tests carried out to evaluate these characteristics are shown in Table 118. Table 118: Details of Tests Conducted Property Test Procedure Noise The tire/pavement noise was measured using the Close Proximity Method (CPX). In this method microphones are placed near the tire/pavement interface to directly measure the tire/pavement noise levels. It was developed in Europe and is defined by ISO Standard 11819-2. CPX testing was also conducted at three different vehicles speeds to evaluate the effect of vehicle speed on tire/pavement related noise. Skid Resistance The wet skid resistance of the pavement surfaces was determined using ASTM E274-97, Standard Test Method for Skid Resistance of Paved Surfaces Using a Full-Scale Tire. The tests were conducted at a vehicle speed of 40 mph using a ribbed tire conforming to ASTM E501-94, Standard Specification for Standard Rib Tire for Pavement Skid-Resistance Tests. Ride Quality Ride Quality measurements were conducted using the NJDOT’s Automatic Road Analyzer (ARAN). Note: The tire/pavement noise, wet skid resistance, and ride quality measurements were all conducted within a 6 month time period of one another. In the case of the ride quality’s RQI and IRI, the parameters were measured at the same time. Each measurement was typically taken at either 0.1 or 0.2 mile intervals. For comparison purposes, the measurements were averaged over the distance of the test section. Bennert et al provided the results of noise level tests, as well as determination of noise gradients in the different types of surfaces. They indicate that the measured noise levels measured at a vehicle speed of 60 mph are used for the main comparison between the different pavement surfaces. A comparison of the thin-lift surfaces (OGFC, Novachip®, SMA, and Microsurfacing) showed that, on average, the OGFC mixes obtained the lowest tire/pavement generated noise levels. The two test sections containing asphalt rubber (AR-OGFC) had the lowest noise levels of all sections tested. The authors note that the AR-OGFC mixes had the finest gradation of the OGFC mixes, perhaps also added to the noise reducing properties. Bennert et al, indicate that, in order of lowest to highest noise producing surface, the rankings were: 1) OGFC mixes, 2) 9.5mm SMA mix, 3) Micro-surfacing, 4) Novachip®, and 5) 12.5mm SMA mix. They mention that a recently (at the time of the study) placed 12.5mm Superpave mix had an unexpectedly low tire/pavement noise of 97.1 dB(A), although the older 12.5mm mixes were almost 2 dB(A) higher. The PCC diamond grind treatment showed tire/pavement noise characteristics comparable to those of the Micro- surfacing and older surfaces of 12.5mm Superpave and Novachip. The PCC pavements generated the largest tire/pavement noise, approximately 10 dB(A) higher than the OGFC mixes.

242 Bennert et al. indicate that the noise data also showed a general increase in tire/pavement noise as the age of the surface type increases. For example, the five year older Novachip surface had a 1.2 dBA higher noise level than the younger section. Similar trends also occurred for the 12.5mm Superpave and the NJDOT MOGFC- 1 mixes. The authors note that the increase in tire/pavement noise may be attributed to both the natural aging/stiffening of the asphalt material, as well as some surface distress that may have been caused by excessive traffic levels. Bennert et al. indicate that the measured noise gradients usually fell within 0.15 to 0.3 dB(A) per mph. The thin-lift material that achieved the lowest noise gradient was the Novachip® and the NJDOT MOGFC-2 mix. In general, the NJDOT MOGFC mixes had noise gradients ranging from 0.15 to 0.16 dB(A) per mph, with the AR-OGFC sections having noise gradients ranging between 0.19 and 0.20. The Micro-surfacing had the largest noise gradient of 0.3 dB(A) per mph with the non-treated PCC pavements yielding the second largest noise gradient. The diamond ground PCC surface had a noise gradient comparable to the HMA mixes. Bennert et al. indicate that the PCC surface with no finish showed the lowest wet skid resistance, whereas the PCC surface with transverse tine and one of the AROGFC courses showed the highest wet skid resistance. The diamond ground PCC section had the third best wet skid resistance. The authors were surprised to see that the 12.5 mm Superpave mixes performed quite well under the wet skid resistance test, out-performing the more open/gap-graded mixes. Using a recommendation that roadway surfaces should have a minimum wet skid resistance ranging from 35 to 40, when using a ribbed test tire, to maintain vehicle safety in wet weather conditions, the authors note that only one of the PCC with no finish would be considered marginally acceptable. Bennert et al indicate that pavement surface treatments that are optimal for selection should have both low tire/pavement noise and high wet skid resistance. Using this criterion, they note that best performing surface in the study was the asphalt rubber OGFC located on I-195 W, while the poorest performing surface was the PCC without treatment. The 12.5 mm Superpave mixes had the second best ranking based on having low tire/pavement noise and high wet skid resistance. Overall, when comparing only the thin-lift mixtures, the OGFC mixtures had the best performance with the Micro-surfacing being the second most optimal surface type. The authors indicate that the 12.5 mm nominal aggregate sized SMA mixed was the worst performing thin-lift surface treatment. Regarding ride quality measurement results, Bennert et al. mentions that the of the 17 test sections evaluated for RQI, only 8 of the sections obtained a rating of Very Good. The remaining sections, except for the 2 transverse tined PCC sections, were rated as Good. They note that it was difficult to determine which surface type provided the “best” RQI rating since the age of the pavement surface, as well as the contractor’s quality control, may have influenced the results. As example, Bennert et al. points out that after 3 years of service, the Novachip® had a RQI rating of 4.47. However, after 8 years, the Novachip® had an RQI of 3.51. Overall, the 12.5mm Superpave mix and the diamond

243 ground PCC had the best RQI ratings, with the OGFC mixes having the best RQI rating for the thin-lift HMA mixes. Since the RQI is derived from the user’s perception of ride quality, Bennert et al hypothesized that the tire/pavement noise may contribute to the user’s rating of the pavement surface. With a plot of RQI versus tire/pavement noise at 60 mph, the authors show that a user’s pavement roughness perception correlated well (R2 = 0.81) to the tire/pavement noise generated during the user’s ride. Bennert et al. also provide the IRI data for the different sections. They note that the IRI values followed a trend which was similar to the trend in the RQI data. They mention that on average, the pavement surface that obtained the smoothest rating (lowest roughness from IRI) was the 12.5mm Superpave mix, with the diamond ground PCC having the second best IRI measurement. The thin-lift surface treatment that had the smoothest rating was the newer Novachip® surface, with the NJOGFC-1 mix being comparable. The roughest pavement surfaces measured in the study were the transverse tined PCC pavements. Bennert et al. also point out a strong correlation exists between the pavement smoothness (or roughness) and the measured tire/pavement generated noise. They mention that this correlation is most likely because the pavement surface macro-texture that is being measured with the ARAN’s lasers for IRI has a direct influence on the tire/pavement noise. In making overall summary, Bennert et al. mention that the pavement surface type that performed the best under the noise, wet skid resistance, and ride quality testing (RQI and IRI) was the 12.5mm Superpave mix, with the diamond ground PCC performing the second best overall. The OGFC mixes performed comparably and was ranked third overall. The worst performing surface type in the study was the PCC with no surface treatment, which obtained on average the highest noise levels, the lowest wet skid resistance, and the poorest ride quality ratings for both the RQI and the IRI measurements. However, in subsequent sections, Bennert et al provide cost information on the different mixes, and from the comparisons, the benefits of OGFC courses become more apparent. This is because, in both examples, the cost of OGFC was found to be significantly lower compared to the alternative courses. In the case of US 24, Union, where MOGFC and the diamond ground PCC performed equally well (overall), the cost of the MOGFC-1 was $4.60/yd2, whereas the cost of the Diamond Grind was estimated at $6.00/yd2. The authors conclude that the MOGFC-1 provided the required reduction in noise and at only two-thirds the cost of the Diamond Grinding, without sacrificing ride quality. Similarly, in the I 195 section in Hamilton, where Novachip, MOGFC-2 and a 12.5mm Superpave mix were the possible alternatives, the MOGFC-2 provided comparable performance at two thirds the cost of the Novachip and one third the cost of the Superpave mix.