Construction and Maintenance Practices for Permeable Friction Courses (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.

43 Very little has been written on the aspect of structural design of PFC mixtures. Fewer than 20 percent of the papers reviewed contained any discussion on the structural design aspects of PFC mixtures. Most of the papers that did discuss structural design dealt with OGFC mixtures and much of that was from outside the United States. In most instances, agencies simply have a standard lift thickness that is placed and the layer is not considered in the structural capacity of the pavement. Oregon probably has had more experience with OGFC mix- tures than any other state in the United States. Moore et al. indicate that the 1993 AASHTO Guide for Design of Pavement Structures and other deflection-based procedures are used for structural designs with open-graded asphalt mixes in Oregon (47, 48). Deflection testing conducted on the Class F mix indi- cated that the deflection reduction was comparable to a dense- graded mix of a similar thickness, thus Oregon has not altered it structural design procedure (48). A minimum thickness of 2 in. (50 mm) has been specified for Oregon’s OGFC class F mix [increased from 1.5 in. (37.5 mm) used in the past, to reduce laydown and compaction problems] (47, 48). A maximum thickness of 4 in. (100 mm) (in two lifts) has been used. In Spain a lift thickness of 1.6 in. (40 mm) has been estab- lished for PFCs (16). The possibility of using thicker lifts has not been considered. However, this thickness is not based on any structural value, but rather on the water absorption capac- ity (ability to store and transport water). Ruiz et al. (16) indi- cate that the water absorption capacity with 1.6 in. (40 mm) is thought sufficient; therefore, the thickness specified in Spain is based upon the volume of potential rainfall. The 1.6 in. thickness also is used in Austria (49). In Italy, though nothing specifically was discussed on structural design, Ranieri stated that a method of selecting the appropriate layer thickness should be based upon rain intensity (50). Ruiz et al. (16) also indicate that the same structural value is assigned to porous asphalt mixtures as other open or semi- open conventional asphalt mixtures such as road bases. When porous asphalt mixtures are used above pavement struc- tures containing cement-treated road bases, an additional 3⁄4 in. (20 mm) of HMA is provided to assist in preventing reflec- tive cracking. Reflective cracking that appears in a PFC will provide an avenue for water to penetrate into the pavement structure, thereby, increasing the potential for pavement deterioration (16). In the Netherlands PFC layers are placed at a thickness of 2 in. (50 mm); this thickness was selected because of the typ- ical rainfall rates experienced (17, 51). According to Van Der Zwan et al. (17) the pavement design methodology in the Netherlands entails designing to prevent classical bottom- up fatigue cracking. When designing pavement thicknesses, dense-graded mixes are assigned a dynamic modulus value of 1.8 million psi (7,500 MN/m2) and the mixture also must meet specific fatigue properties. Van Der Zwan et al. (17) pro- vided a discussion on the structural contribution of a PFC layer within pavements by comparing dense-graded and PFC. One area that dense-graded and PFC were compared was in terms of dynamic modulus. Van Der Zwan et al. (17) indicated that the dynamic modulus of PFC is generally 783,000 psi (5,400 MN/m2), or about 70 to 80 percent of dense-graded mixes. This value of dynamic modulus was input into their pavement design models and the results indicated that 10 to 20 percent more thickness was required to maintain a specific fatigue strain at the bottom of the pavement layer when using PFC as compared to dense-graded mixes. Van Der Zwan et al. (17) also evaluated the effect of aging and stripping on pavement design. Due to the open nature of PFC, the asphalt binder coating aggregates is susceptible to accelerated oxidative aging. Oxidative aging of the asphalt binder results in an increase in stiffness within the PFC layer which would reduce pavement thickness. Alternatively, the authors state that water within the PFC layer can lead to a loss of adhesion between the PFC layer and the underlying layer. This loss of adhesion impairs the load transfer char- acteristics of the structure. Van Der Zwan et al. (17) state that there is no evidence that the loss of adhesion between layers C H A P T E R 6 Inclusion in Structural Design

(delamination) has taken place in the field; they conserva- tively assumed a loss of adhesion to evaluate the net effect on pavement structure using the BISAR program. When delam- ination occurs, the effective bearing capacity of the debonded layer is reduced to between two and 10 percent of the origi- nal value. By applying Miner’s modified linear damage law, the authors stated that the combined effect of aging and stripping (delamination) would result in about 35 to 40 percent effective contribution of PFC when compared to dense-graded layers. Because PFC has different thermal properties than dense- graded mixes, Van Der Zwan et al. (17) also evaluated the effect of temperature on the pavement structure when comparing dense-graded and PFC wearing layers. Van Der Zwan et al. (17) provided a hypothesis that the suction and pumping action of tires passing over PFC surfaces, coupled with wind action, promotes continuous air circulation within a PFC layer. As a result, the temperature of the PFC layer will tend to be lower than for comparable dense-graded layers. In order to investi- gate this hypothesis, the authors conducted experiments on newly placed and 8-year-old PFC layers to compare the tem- peratures of pavements with porous asphalt and dense-graded wearing layers at the surface and at depth. Results from these experiments, which included one year of data, indicated that the weighted average temperature over a year was found to be 1.8°F (1°C) lower in pavements containing PFC wearing layers. Due to the viscoelastic properties of asphalt, the lower effective temperature in pavements including a PFC wearing layer, means that the stiffness (modulus) of underlying layers is higher. The net result shows that less thickness is required to resist fatigue cracking. The combined effect of these factors suggests that PFC can be expected to contribute about 50 per- cent of the equivalent structural capacity compared to a dense- graded layer. However, if adhesion between the PFC layer and the underlying layer is not lost (as was conservatively assumed), then the effective contribution of porous asphalt can be 100 to 110 percent of conventional systems (17). Watson et al. (52) also concluded that layers underlying open-graded wearing layers are cooler than mixes underlying typical dense-graded wearing layers. These conclusions are based upon temperature measurements made at the pavement surface and at depth on the 2000 National Center for Asphalt Technology Test Track. Layers underlying open-grade surfaces averaged about 3.8°F (2°C) cooler than pavement layers underlying conventional dense-graded layers. Similar to the conclusions of Van Der Zwan et al. (17), this would result in an increase in stiffness for underlying layers which would, there- fore, improve the structural capacity of underlying layers. In Switzerland, the typical layer thickness for PFCs ranges from 1.1 to 2 in. (28 to 50 mm) (15). According to Isenring et al. (15) PFC mixes in Switzerland having a maximum aggre- gate size of 1⁄2 in. (10 mm) are typically placed 1.1 to 1.7 in. (28 to 42 mm) thick while porous asphalt mixes having a max- imum aggregate size of 0.625 in. (16 mm) are typically placed from 1.75 to 2 in. (43 to 50 mm) thick. The British Columbia Ministry of Transportation and High- ways currently considers a structural strength value of 1.25 (in terms of Crushed Granular Equivalency) for OGFC. This can be compared to 2.0 for conventional asphalt pavement (53) or about 60 percent of the structural value of dense-graded mixes. According to Van Heystraeten and Moraux (22), in Belgium two thicknesses are used with PFCs: 1 and 1.6 in. (25 and 40 mm). To maintain the drainage characteristics and noise- reducing attributes for a longer period of time, they indicate that the 1.6 in. layer thickness is best. They also stated that based upon modulus testing, PFCs constructed with an 80/100 pene- tration graded asphalt binder will contribute 73 to 79 percent of the structural capacity of typical dense-graded mixes (22). In their paper, Bolzan et al. (26) discussed some structural considerations of PFCs layers. On the basis of the facts that PFC is a mixture in which fractions of the aggregate grad- ing are absent, they contend that these mixtures have lower strength than dense-graded mixtures. They mention that some researchers accept that these mixtures have up to 70 per- cent of the strength of a conventional mixture; others indi- cate the ratio is only 50 percent while the Spanish believe that they are structurally equivalent with conventional dense- graded asphalt mixtures. They also are considered to be less shear stress resistant. Bolzan et al. (26) indicate that Argentina adopted a 50 percent structural capacity for PFC mixtures in the initial projects. The resilient modulus (ASTM D 4123) at 77°F (25°C) and 10 Hz of PFC mixtures, prepared in the laboratory of Argentina, was found to be about 319,000 psi (2200 MPa), approximately 60 percent of the conventional mixtures. However, Bolzan et al. point out that at both higher and lower temperatures, polymer-modified PFC mixes per- form better than unmodified conventional mixes and that further research needed to be conducted to reach any defin- itive conclusions (26). Within the United States, there has recently been a move toward mechanistic-empirical pavement design methods. Within this new pavement design system, HMA mixtures are characterized using the dynamic modulus (E*) test. A single reference was identified that looked at measuring E* for open- graded mixes. Kaloush et al. (54) conducted a study to evalu- ate the E* of asphalt-rubber mixes in Arizona. Of the two dif- ferent asphalt-rubber mixes evaluated, one was labeled as a Asphalt Rubber Asphalt Concrete-Gap Grade (ARAC) mix and one an Asphalt Rubber Asphalt Concrete Friction Course- Open-Graded (AR-ACFC) mix. Kaloush et al. (54) stated that the AR-ACFC had in-place air voids of 18 percent; therefore, it is considered a PFC. Within the research, the authors evalu- ated various confining pressures including 0, 10, 20 and 30 psi 44

(0, 69, 138, and 207 kPa, respectively). Kaloush et al. (54) concluded that the confining pressure used during testing was important. Using E* values from typical Arizona DOT dense-graded mixes and the asphalt-rubber mixes at 20 psi (138 pKa) confinement, the authors showed that the asphalt- rubber mixes had lower modulus values at low temperatures and higher modulus values at high temperatures. Therefore, asphalt-rubber mixes were ranked above the conventional mixes in terms of both low and high temperature performance. Citing these results, Kaloush et al. (54) indicated that confined test results of E* are more appropriate for evaluating gap- and open-graded mixes. From the web-based survey conducted by the research team, it was learned that only seven states assigned a structural value to OGFC pavements. Most of the states that did assign a struc- tural value used a layer coefficient for estimating structural value. Only Texas used a resilient modulus, but limits the use of PFC mixtures to pavements that are already structurally sound. California does not consider the structural benefits of PFC when determining layer thicknesses for the structural sec- tions even though a structural value is assigned. Rational Method of Selecting PFC Lift Thickness Within most of the United States, the thickness of PFC lay- ers placed on the roadway has been based upon experience. No formalized method of determining an appropriate lift thickness was found in the literature. Therefore, a method for selecting an appropriate lift thickness for PFCs was developed. As stated previously, NCHRP Project 9-41 did not include any laboratory or field experiments, and therefore this method has not been validated. Historically, a single lift thickness of PFC has been speci- fied by a given state agency. For instance, Oregon has long utilized 2 in. (50 mm) as a standard lift thickness. Likewise, Georgia specified 3⁄4 in. (19 mm) for a number of years when utilizing OGFCs and then changed to a standard lift thickness of 11⁄4 in. (32 mm) for PFCs. Selection of a standard lift thick- ness of PFC by a state agency that is based upon experience is valid. However, for agencies that have not utilized PFC or only have limited experience with PFCs, a systematic and practical method of determining an appropriate lift thickness is needed. This section provides a practice for selecting appro- priate lift thickness for PFC layers. In order to develop a method for selecting an appropriate lift thickness of PFC, the first question that must be asked is, “What attributes of a PFC layer are desirable?” The literature review and survey provided numerous benefits of PFC. Pre- dominately, these benefits were related to the ability of PFC to drain water during a rain event. Therefore, permeability should be a parameter that is considered during selection of an appro- priate lift thickness. Ruiz et al. (16) state this property was the criteria utilized in Spain for selecting a lift thickness. Tan et al. (55) and Ranieri (50) also deemed permeability as an impor- tant property for determining an appropriate lift thickness. Obviously, some measure of precipitation for the project loca- tion also is needed. Other parameters that would be needed to define the ability of a PFC layer to drain water also would include geometric properties of the pavement section such as lane width and cross-slope. All of these parameters discussed are known or can be readily obtained at the time of pavement design, except for permeability. Based upon the literature, most, if not all, measures of true PFC permeability have been from labora- tory testing. Field tests described in the literature only pro- vide an index of permeability. Unfortunately, the labora- tory tests only estimate a vertical coefficient of permeability with no horizontal component. Therefore, an assumption has to be made that the results of laboratory permeability tests are representative of the in-place permeability of PFC and that the coefficient of permeability is uniform in three dimensions. Methodology Ranieri (50) has indicated that the flow of water through a PFC layer is similar to the flow of water through an uncon- fined aquifer. This analogy seems appropriate for PFCs. Un- confined aquifers generally have an impermeable layer beneath the aquifer and a free surface of water at the top (Figure 22). The underlying impermeable layer within the pavement sys- tem would be represented by a dense-graded HMA along with an appropriately applied tack coat. As long as the PFC is a wearing surface, water residing within the PFC layer would 45 hL hmaxho d L x I I I II Recharge Rate Figure 22. Flow through and unconfined aquifer with constant recharge.

be considered a free surface since the water would be open to atmospheric pressure. An applicable model utilized in groundwater hydrology for unconfined aquifers is the Dupuit equation. Dupuit developed his model based upon the following assumptions (56): • The free surface is only slightly inclined. • Streamlines may be considered horizontal and equipotential lines vertical. • Slopes of the free surface and hydraulic gradient are equal. Dupuit utilized Darcy’s law for one-dimensional flow per unit width and an unconfined aquifer to derive Equation 1. This equation is used to calculate the flow of water through the unconfined aquifer based upon varying heads of water on either side of the aquifer (Figure 22) (56). where q = flow of water through the unconfined aquifer; K = coefficient of permeability; L = length of flow path; ho = upper hydraulic head of water; and hL = lower hydraulic head of water. Dupuit’s equation shown in Equation 1 does not take into account recharge of an aquifer. For the case of a PFC layer, a rain event would represent recharge. When an unconfined aquifer is recharged, the free surface of water takes the form of a parabola as shown in Figure 22 (56). This figure also shows the rate of recharge (I), locations of specific hydraulic heads (ho, hL and Hmax), and the distance to the water divide (peak of parabola) (shown as distance d in Figure 22). Dupuit utilized the information shown in Figure 22 and Equation 1 to derive an equation for the shape of the parabola (Equation 2). q K L h ho L= −( ) 2 2 2 Equation 1 where x is distance as shown in Figure 22. At the peak of the parabola (distance d), a boundary condi- tion exists such that no flow occurs (56). Therefore, Equation 2 will take the form of Equation 3 at x = d and q = 0 (56). Figure 23 illustrates the applicability of this approach to a PFC layer having a unit width. The PFC is assumed to be over- lying an impermeable layer, which would be a combination of a dense-graded HMA and tack coat. There are two hydraulic heads acting on the flow of water through the layer. The upper hydraulic head (ho) would be equal to the length (L) times cross slope (α) plus thickness (t) of the PFC layer. The lower hy- draulic head (h2) would essentially be zero. A rain event would represent the recharge rate. Within Equation 3, there is one un- known that must as assumed, the distance to the peak of the parabola (d). Being a parabola, a reasonable assumption would be that the distance to the peak of the parabola is equal to one- third of the length (L/3). Using this simple assumption, Equa- tion 3 can be solved to determine the needed thickness of PFC to prevent a sheet of water from developing at the location of hmax. Thickness is contained within the ho term of Equation 3. Likely the most important property required to determine the required thickness of a PFC layer is a measure of rain inten- sity. As rain intensity increases, more thickness would be needed to store and move the water to the pavement edge. An easily accessible database for obtaining a measure of rain inten- sity is contained on the National Oceanic and Atmospheric Administration (NOAA) web page at http://cdo.ncdc.noaa. gov/cgi-bin/HPD/HPDStats.pl from a link entitled, “Hourly Binned Precipitation in Microsoft Excel Format,” as shown in Figure 24. Any designer can download this Microsoft Excel 0 2 2 2 2 = −( )+ −⎛⎝⎜ ⎞⎠⎟ K L h h I d L o L Equation 3 q K L h h I x L o L= −( )+ −⎛⎝⎜ ⎞⎠⎟2 22 2 Equation 2 46 L = Length, ft X t = Thickness, in. hL ho Cross slope (α), % Pavement Edge Crown Rainfall Intensity (I), inches per hour Figure 23. PFC layer simulating an unconfined aquifer with constant recharge.

database onto a computer. This database (hereafter called the Hourly Precipitation Database), a portion of which is shown in Figure 25, contains hourly precipitation statistics for over 6,000 weather stations. Each weather station is coded within the database by its “COOP ID.” The COOP ID can easily be found from the above referenced web page shown in Figure 24. Simply select the state for which the project is to be located and a list of available weather stations will come up. From this list, the designer can select the closest weather sta- tion to the project site and obtain the COOP ID station name. Then, within the Microsoft Excel database simply conduct a find query from the “Edit” menu on the COOP ID station number. Alternatively, the COOP ID values are in ascending order within the Hourly Precipitation Database. Therefore, a designer should have no problem finding the specific weather station data within the Hourly Precipitation Database. Figure 25 illustrates a portion of the data from the Hourly Precipitation Database. The first column is the COOP ID num- ber, and the second and third columns provide the beginning and ending years when data is available for the selected weather station. The fourth column provides the number of days a rain event(s) occurred during the time that data is available. The next two columns provide the number of days and per- cent of total days the suspect data was obtained. Following the flagged observation data are alternating columns of the number of rain events and percentage of those rain events for various increments of rain intensity in inch per hour units. A breakdown of the data is provided in 0.1 in. increments up to 1.1 in. per hour, then from 1.1 to 1.5 in. per hour and greater than or equal to 1.6 in. per hour. As an example, the hourly precipitation rainfall statistics were obtained for Canton, Mississippi. Using the web page shown in Figure 24, the State of Mississippi was selected. Of the available weather stations in Mississippi, the Canton station was identified and had a COOP ID of 221389. Next, this COOP ID number was found in the Hourly Precipitation Database. Hourly rainfall data obtained from the database are provided in Table 29. This table shows the increments of rain inten- sity, the percentage of rain events within each increment, and the cumulative percentage of rain events from the different increments. Using the larger number of each increment shown in Table 29 and the cumulative percentage within each increment, a chart can be developed that describes the percentage of all rain events below a given intensity (Figure 26). A designer can use the information in Figure 26 to select rainfall intensity for use in selecting the PFC lift thickness that is representative of the project location. In Figure 26, rainfall intensity at 90 percent of all rain events (I90) was approximately 0.4 in. per hour. As stated previously, other information for determining a minimum PFC lift thickness includes the design cross slope (α) of the roadway as well as the length of drainage path (L). The length of the drainage path would be the length of PFC from the highest point to the lowest, in a transverse direction. 47 Database Figure 24. NOAA web page with hourly precipitation database.

48 CO O P ID Be gi n Y ea r En d Ye ar # o f D a ys # o f D a ys w /fl a gs % F la gg e d > = . 01 < . 1 Ev en ts > = .0 1 < . 1 % O f T o ta l > = .1 < .2 Ev en ts > = .1 < .2 % O f T o ta l > = .2 < .3 Ev en ts > = .2 < .3 % O f T o ta l > = .3 < .4 Ev en ts > = .3 < .4 % O f T o ta l > = .4 < .5 Ev en ts > = .4 < .5 % O f T o ta l > = . 5 < . 6 Ev en ts > = .5 < .6 % O f T o ta l > = . 6 < . 7 Ev e n ts 010008 1948 2002 5206 695 13.35 8078 48.98 4762 28.87 1579 9.57 746 4.52 423 2.56 273 1.66 159 010063 1948 2002 5404 530 9.81 9766 48.56 6271 31.18 1942 9.66 894 4.45 429 2.13 262 1.3 154 010140 1963 2002 3903 143 3.66 1677 15.19 5700 51.63 1615 14.63 800 7.25 410 3.71 249 2.26 187 010252 1980 2002 2145 249 11.61 0 2955 59.49 865 17.41 423 8.52 209 4.21 137 2.76 96 010272 1948 2002 877 738 84.15 1125 73.87 199 13.07 81 5.32 58 3.81 22 1.44 16 1.05 5 010369 1948 2002 5464 502 9.19 8317 43.35 6611 34.46 1939 10.11 928 4.84 520 2.71 293 1.53 161 010390 1982 2002 1804 264 14.63 0 2823 63.85 791 17.89 358 8.1 196 4.43 88 1.99 43 010402 1948 2002 3407 292 8.57 4622 42.73 3453 31.92 1112 10.28 547 5.06 292 2.7 187 1.73 144 010407 1965 1982 1626 125 7.69 1987 37.41 1859 35 568 10.69 313 5.89 174 3.28 106 2 75 010425 2000 2002 133 42 31.58 0 135 61.93 45 20.64 16 7.34 10 4.59 4 1.83 2 010427 1948 1962 1320 41 3.11 3024 57.21 1150 21.76 489 9.25 254 4.81 129 2.44 82 1.55 47 010430 1979 1996 1758 63 3.58 1667 29.93 2402 43.13 703 12.62 329 5.91 159 2.86 103 1.85 69 010631 1948 1955 836 22 2.63 1968 61.1 646 20.06 278 8.63 142 4.41 73 2.27 40 1.24 29 010748 1948 2002 4855 654 13.47 7591 46.48 5359 32.81 1514 9.27 752 4.6 390 2.39 231 1.41 157 010790 1948 1967 1922 221 11.5 6193 73.04 1206 14.22 519 6.12 207 2.44 111 1.31 67 0.79 48 010829 1978 1990 1393 2 0.14 4316 71.08 897 14.77 375 6.18 190 3.13 99 1.63 69 1.14 29 010831 1948 2002 5413 99 1.83 12692 60.36 4760 22.64 1605 7.63 779 3.7 414 1.97 243 1.16 169 010836 1949 1949 8 0 0 19 61.29 6 19.35 3 9.68 0 1 3.23 0 010957 1948 2002 5053 481 9.52 10075 52.4 5705 29.67 1738 9.04 737 3.83 352 1.83 207 1.08 118 011099 1982 2002 1983 210 10.59 0 3494 64.97 973 18.09 420 7.81 200 3.72 109 2.03 70 011315 1948 1967 1773 447 25.21 4276 70.54 860 14.19 388 6.4 173 2.85 116 1.91 72 1.19 42 011819 1951 1980 3343 106 3.17 10464 69.03 2459 16.22 1056 6.97 456 3.01 277 1.83 162 1.07 79 012124 1948 2002 5495 279 5.08 7243 38.62 7127 38 2028 10.81 987 5.26 505 2.69 300 1.6 164 012172 1975 2002 2535 115 4.54 0 3865 57.23 1167 17.28 558 8.26 321 4.75 222 3.29 148 Figure 25. Example of information available in hourly precipitation database. Increment of Rainfall Intensity (in./hr) Percent Events Within Increment Cumulative Percentage >=.01 <.1 % Of Total 31.3 31.3 >=.1 <.2 % Of Total 41.78 73.1 >=.2 <.3 % Of Total 12.39 85.5 >=.3 <.4 % Of Total 5.63 91.1 >=.4 <.5 % Of Total 2.9 94 >=.5 <.6 % Of Total 1.83 95.8 >=.6 <.7 % Of Total 1.25 97.1 >=.7 <.8 % Of Total 0.82 97.9 >=.8 <.9 % Of Total 0.54 98.4 >=.9 <1.0 % Of Total 0.38 98.8 >=1.0 <1.1 % Of Total 0.28 99.1 >=1.1-1.5 % Of Total 0.63 99.7 >=1.6 % Of Total 0.27 100 Table 29. Rainfall intensity data for Canton, Mississippi. For two-lane highways with a crown at the centerline, this value would generally be 12 ft (3.6 m). For four-lane divided highways with no crown at the center of the two lanes, this value would be 24 ft (7.3 m). Figure 26 illustrates the cumulative percentage of rain events versus rain intensity for Canton, Mississippi. Within the exam- ple problem, the rainfall intensity at 90 percent of all rain events was identified. The pavement designer must select the cumu- lative percentage of rain events from which rain intensity is selected. Based upon the relationship between percentage of rain events and rain intensity, a cumulative percentage of 90 percent (I90) seems appropriate for selecting rain intensity. However, a pavement designer can utilize any percentage of rain events desired. The final piece of information required is an estimate for per- meability. This estimate can come from the design of the PFC mixture or assumed as 328 ft/day (100 m/day). Mallick et al. (36) have recommended a minimum permeability of 328 ft/day

(100 m/day); therefore, this assumption will be conservative. Also, agencies that have conducted research on PFC may have other estimates of permeability for their materials that can be used. This data described here and Equation 3 was utilized to develop a series of design curves that a designer could use to select the desired PFC lift thickness (Figure 27). The design curves shown in Figure 27 are based upon different design cross slopes. Obviously, the slope of a pavement is ever changing due to horizontal and vertical curves. The cross slope refer- enced in Figure 27 represents the cross slope in a flat region of roadway (not within a horizontal or vertical curve). 49 Hourly Precipitation Data Percent of Total Precipitation Events 0 10 20 30 40 50 60 70 80 90 100 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.60 Rain Intensity, in/hr Pe rc en t o f R ai n Ev en ts , % 0.4 in/hr≈ Figure 26. Cumulative percentage of rain events versus rain intensity. Design Chart for PFC Lift Thickness Based on Cross Slope 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 0.000 0.004 0.008 0.012 0.016 0.020 0.024 0.028 0.032 0.036 0.040 0.044 4I/K t/L Cross Slope = 2.5% Cross Slope = 2.0% Cross Slope = 3.0% Cross Slope = 3.5% Cross Slope = 4.0% Cross Slope = 1.5% Figure 27. Design curves for selecting PFC layer thickness.

To use the design curves shown in Figure 27, the required data previously described must first be obtained. Table 30 pre- sents the required data, along with the applicable units of each property, for using the design curves. Results from the use of the design curves are a PFC layer thickness in the units of millimeters. In order to show the usefulness of the design curves, three examples are provided for Miami, Florida; Atlanta, Georgia; and Las Vegas, Nevada. For each example, permeability (K) is assumed as 100 m/day (328 ft/day), cross slope (α) of 2.0 percent, and a drainage path (L) of 12 ft (3.6 m). As stated previously, the first step of the process is to deter- mine the rain intensity values for each location. To obtain this data, the NOAA web page and Hourly Precipitation Database were utilized. Figure 28 presents the rain intensity versus per- centage of rain events for all three locations. Based on this fig- ure the rain intensity at a level of 90 percent of all rain events (I90) would be as follows: Location I90 (in./hr) Miami, Florida 0.37 Atlanta, Georgia 0.22 Las Vegas, Nevada 0.10 Using the rain intensity data and the assumptions, the designer can go to the design curve chart for a cross slope of 2 percent to determine the required PFC lift thickness. Table 31 and Figure 29 illustrate the progression for determining the required PFC lift thickness. Table 31 also shows the determined lift thicknesses (t). This example points out that there has to be a practical min- imum lift thickness for PFC. In the example, the required thick- ness of PFC in Las Vegas could not be determined because of the relatively insignificant rainfall intensity (I90) observed there. From a practical standpoint, a minimum layer thickness should be based upon the maximum aggregate size of the PFC grada- tion. Lift thicknesses should likely be at least 1.5 to 2 times the maximum aggregate size of the gradation. There is limited evidence that thicker PFC layers that have a relatively high level of permeability can maintain permeability for a longer period of time, that is, less clogging. Sensitivity Analysis A sensitivity analysis was conducted to evaluate the devel- oped method of determining a PFC lift thickness. In con- ducting the sensitivity analysis, reasonable input values were assumed in order to evaluate their effect on the resulting layer thickness. As described previously, there are a total of four inputs required for the developed methodology which include: a measure of permeability, rain intensity, design cross-slope, and the length of the flow path (width of pavement from high point to low point). 50 UnitsRequired Properties m/dayPermeability, K in./hrRainfall Intensity, I mLength, L Table 30. Required properties for design curves. Use of Design Curve Example Selection of Rain Intensity (I) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00 Rain Intensity (I), in/hr Pe rc en t o f R ai n Ev en ts , % Miami, Florida Atlanta, Georgia Las Vegas, Nevada Figure 28. Rain intensity versus cumulative percent rain events for example problem.

51 Design Chart for PFC Lift Thickness Based on Cross Slope 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 0.020 0.022 0.024 0.026 0.028 0.030 0.032 0.034 0.036 0.038 0.040 0.042 4I/K t/L Cross Slope = 2.5% Cross Slope = 2.0% Cross Slope = 3.0% Cross Slope = 3.5% Cross Slope = 4.0% Cross Slope = 1.5% Miami Las Vegas Atlanta Figure 29. Example problem determining PFC layer thickness. Location K, m/day I, in./hr (Figure 31) 4I/K t/L (Figure 32) L,m t,mm Miami 100 0.37 0.015 11.1 3.6 40.0 Atlanta 100 0.22 0.009 4.0 3.6 14.4 Las Vegas 100 0.10 0.004 --- 3.6 --- Table 31. PFC lift thickness example based upon 2 percent cross slope. Figure 30 illustrates the effect of permeability on the result- ing lift thickness. For this figure, the rainfall intensity was set at 0.4 inches per hour, cross-slope was set at 2 percent, and the flow path was set at 12 ft (3.6 m). Based upon the figure, as permeability increases the needed thickness decreases, which seems logical. As permeability of the layer increases, the abil- ity of the layer to drain water to the pavement edge increases. Therefore, less volume of water is stored within the PFC at any given moment (assuming a constant rain intensity) lead- ing to less needed thickness. The range of permeability values presented in Figure 30 is 100 to 250 m/day. Within this range of permeability values, Figure 30 shows a significant effect of permeability on the needed PFC layer thickness. This indicates that selection of an appropriate permeability value is very important in using the proposed methodology. Figure 31 illustrates the influence of rain intensity of the resulting layer thickness. For this figure, the flow path length was 12 ft (3.6 m), cross slope was 2 percent and permeability was 328 ft/day (100 m/day). As rain intensity increases, the needed thickness of PFC also increases. Rainfall intensities included within Figure 31 ranged from 0.25 to 0.5 in./hr. Except within the drier portions of the country, this range appears typical. The effect of cross slope on the needed PFC layer thickness is shown in Figure 32. As cross slope increases, the needed PFC layer thickness decreases. As cross slope is increased for a given lane width, the hydraulic head also increases which helps to drive the water within the PFC to the pavement edge and, thus, discharge. Within Figure 32, rainfall intensity, I90, was set at 0.5 in. per hour, permeability was set at 328 ft/day (100 m/day) and the length of flow path was 12 ft (3.6 m). Figure 32 (as well as Figure 27) provides an interesting oppor- tunity for pavement designers. In areas susceptible to very intense rain events or multiple lanes (longer flow path), sim- ply increasing the design cross slope by 0.5 percent can signif- icantly reduce the needed thickness of PFC. The final input into the PFC layer thickness methodology is the flow path length. Flow path length will change as the

number of lanes to be paved with PFC changes. Also, extend- ing the PFC the full width of the shoulder will increase the flow path length. Figure 33 illustrates the effect of flow path length on the design lift thickness. For this figure, rainfall intensity (I90) was set at 0.3 in. per hour, the cross slope was set at 2 per- cent and permeability was set at 328 ft/day (100 m/day). As the flow path length increases, the required thickness of PFC also increases because the longer flow path means that the water is held in the PFC longer. Therefore, more thickness is required to store the water while it drains to the pavement edge. For the data shown in Figure 33, an increase in flow path length from 12 ft (3.6 m) to 24 ft (7.6 m) almost doubled the required thickness of PFC. These discussions indicate that each of the four input param- eters have a significant effect on the resulting thickness of a PFC layer when using the methodology described in this chapter. Therefore, the input values must be carefully selected. Of the four factors, the rainfall intensity (I90) likely has the 52 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 Rain Intensity, inch per hour PF C La ye r T hi ck ne ss , m m Length, L = 3.6 m Cross Slope, = 2 percent Permeability, K = 100 m/day Figure 31. Effect of rain intensity on design lift thickness. 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 90.0 110.0 130.0 150.0 170.0 190.0 210.0 230.0 250.0 270.0 Permeability, m/day PF C La ye r Th ic kn es s, m m Rainfall Intensity, I = 0.4 inches per hour Cross Slope, = 2 percent Length, L = 3.6 m Figure 30. Effect of PFC permeability on design lift thickness.

most influence on thickness; however, as described, the I90 values are generally in a very narrow range. Discussion of Proposed Methodology A number of approaches were investigated to develop a methodology for determining a minimum layer thickness of PFC. Of the different approaches, the use of principles associated with an unconfined aquifer was the most logical and practical. The term practical, as used here, means that the resulting minimum lift thicknesses were realistic with current practices. Intuitively, the use of concepts associated with uncon- fined aquifers is very similar to the flow of water through a PFC layer. Unconfined aquifers have an impermeable layer under- neath and a free water surface at the top. 53 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 Cross Slope, % PF C La ye r T hi ck ne ss , m m Rainfall Intensity, I = 0.5 inches per hour Permeability, K = 100 m/day Length, L = 3.6 m Figure 32. Effect of cross slope on design lift thickness. 0.0 10.0 20.0 30.0 40.0 50.0 60.0 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Length, m PF C La ye r T hi ck ne ss , m m Rainfall Intensity, I = 0.3 inches per hour Cross Slope, = 2 percent Permeability, K = 100 m/day Figure 33. Effect of flow path length on design lift thickness.

The example problem presented earlier indicates that the results from the described approach are realistic. Miami, Florida receives a significant amount of rainfall compared to much of the United States and based upon the design curves and assumptions made, an appropriate thickness is 40 mm. This thickness seems appropriate and is in line with thick- nesses used around the United States and in Europe. There- fore, the minimum lift thickness approach passes the test of reasonableness for the Miami example. The results of the example problem for Atlanta, Georgia were a PFC thickness of 14.4 mm. This value is actually less than the typical PFC thickness used in Georgia of 32 mm. However, using a thickness layer larger than that determined using the design curves should not be discouraged. As stated previously, there is evidence that thicker layers of PFC main- tain permeability longer. The action of traffic and the larger volume of voids (not percentage but overall volume) help clean the void structure through hydraulic action of tires passing over the layer. The one potential problem with the proposed design curves has to do with permeability. No matter how well general main- tenance is performed, there always will be some degree of clog- ging within a PFC layer. If the permeability of the PFC layer becomes lower than the assumed permeability value, theo- retically, a sheet of water could exist at the pavement surface. This is not perceived as a problem, however, because the very high macrotexture of a PFC pavement should more than offset any film of water that exists on the pavement surface. The relatively high macrotexture of PFC wearing layers pro- vides channels for water to be displaced when a tire passes over the pavement surface. Additionally, the design curves do not take into account any hydraulic action caused by traffic which results in water being “pushed through” the void structure of the PFC layer, which makes the design curves conservative. 54