Practices for Unbound Aggregate Pavement Layers (2013)

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.

40 chapter three GRANULAR BASE AND SUBBASE CONSTRUCTION PRACTICES INTRODUCTION Aggregate storage, transportation, and construction practices are critical to ensuring adequate performance of constructed UAB and subbase layers under loading. Improper material handling and construction procedures often lead to aggre- gate segregation and/or degradation, ultimately resulting in a poorly compacted aggregate layer. Because unbound aggre- gate layers function primarily through interparticle load trans- mission at aggregate contact points, such poorly compacted layers may undergo excessive shear deformation, leading to pavement failure. This chapter comprises an overview of common construc- tion and material handling practices adopted by transportation agencies as far as UAB and subbase layers are concerned. Extensive review of published literature was conducted to identify different methods identified by researchers in the past as being adequate or inadequate for aggregate base and sub- base construction. A survey of U.S. state and Canadian pro- vincial agencies was conducted to gather information on the state of the practice on this topic, and an analysis of the find- ings presented to highlight areas where significant improve- ments are still needed. Moreover, applications of nonstandard or unconventional pavement types using unbound aggregate layers and related construction practices, such as the inverted pavement concept of a granular layer over a stiff layer at depth, are described in this chapter. The overall objective is to identify gaps in knowledge concerning the “effective prac- tices” for UAB and subbase layer construction, along with research needs to address these gaps. IMPORTANCE OF STANDARDIZED CONSTRUCTION SPECIFICATIONS A “pocket-sized” handbook published by the NSSGA (1989) contains important guidelines for UAB construction. Simi- larly, different transportation agencies have adopted different guidelines to help the construction of “good quality” base and subbase layers. Apart from providing the contractors with a definite set of guidelines to be followed during construction, these guidelines help the field engineers with QA of con- structed pavement layers. However, the survey of state and Canadian provincial transportation agencies conducted under the scope of this synthesis study indicated that only 37% of the responding agencies (17 of 46) currently have specific guidelines regarding the transportation and storage (stock- piling) of aggregate materials for base and subbase construc- tion. Approximately 25 agencies reported not having any such guidelines, whereas the remaining four agencies indicated the presence of generic guidelines without specific instructions. AGGREGATE STORAGE AND CONSTRUCTION PRACTICES AFFECTING CONSTRUCTED LAYER PERFORMANCE To fulfill the overall objectives of this synthesis study, it is important to first present a summary of material handling and construction practices that have been identified as adequate or inadequate as far as ensuring the construction of good qual- ity unbound aggregate pavement layers. Inadequate material handling and construction practices may lead to aggregate seg- regation and/or degradation affecting the gradation or particle size distribution of the constructed aggregate layer. The fol- lowing sections discuss different material storage and con- struction practices that may lead to the problems of aggregate segregation and degradation. Aggregate Stockpiling as a Source of Segregation The Aggregate Handbook defines aggregate segregation as the separation of one size of particles from a mass of par- ticles of different sizes, caused by the methods used to mix, transport, handle or store the aggregate in the plant under conditions favoring nonrandom distribution of the aggregate sizes (Barksdale 1991). Certain practices magnify the segre- gation problem and thus are best restricted by transportation agencies. One possible source of segregation is during the formation of conical stockpiles by dumping material using a conveyor belt. As the aggregate is transported by a conveyor belt, vibration and motion of the belt causes the fine particles to settle to the bottom of the material stream, whereas coarse particles remain at the top. These coarse particles have a higher velocity at the end of the conveyor, and are thrown a greater distance to the stockpile. In addition, the coarser par- ticles hit the front face of the stockpile with a greater momen- tum and roll down the outer edge of the pile, creating overrun (an accumulation of particles at the pile’s bottom edge or toe). Fine particles, which have settled against the surface of the conveyor belt, tend to cling to the belt and drop to the back face of the pile. The resulting stockpile is segregated, with coarse particles settled at the toes, and fine particles in the center portion of the pile.

41 “Material overrun,” particles (regardless of size) moving down the side of the stockpile, is another major source of seg- regation in stockpiles. As the material moves down the side of the stockpile, larger particles tend to move down to the bottom (owing to higher momentum), whereas finer materials tend to settle into the side of the pile. Such spatial distribution of aggregate particles of different sizes at different portions of the stockpile results in pronounced segregation. Figure 24 shows the spatial distribution of different aggregate particle sizes in a segregated stockpile (Nohl and Domnick 2000). Materials with a large variation in particle size usually undergo higher degrees of segregation as a result of improper stockpiling practices. Usually aggregate materials in which the ratio of the largest to the smallest particle size exceeds 2:1 are likely to experience segregation problems during stock- piling (Nohl and Domnick 2000). From in-depth investigation of aggregate stockpiling practices, Miller Warden Associates (1964) observed that flat-mixed piles formed by the use of a crane bucket was the only stockpiling method that resulted in an insignificant amount of segregation. The most commonly used truck dumping method, although economical, was found to cause significant segregation of aggregates. Majidzadeh and Brahma (1969) studied different stages in the aggregate handling process, such as (1) initial material fabrication, (2) producer stockpile, (3) truck transportation, and (4) job- site stockpile, to establish the severity of segregation problem at these different stages and also observed that the segrega- tion problem increased as the material approached the job site from the production plant. Creating stockpiles using the “windrow concept” is one of the alternatives available for storage of materials where segre- gation is a likely problem. Involving the creation of “miniature stockpiles” in layers, windrow stockpiles can be built effec- tively using a telescoping conveyor that can move laterally as well as along the direction of the conveyor to create the stockpile in layers. Although individual “miniature stockpiles” in a windrow pile still undergo segregation, such stockpiles are said to have better “segregation resolution” because the seg- regation pattern repeats itself in smaller intervals. Figure 25 shows schematics of (a) the configuration of a windrow pile formed using a telescoping conveyor and (b) segregation reso- lution in a windrow pile (Nohl and Domnick 2000). Stockpiling Practices by Different Agencies Different transportation agencies adopt different stockpiling practices to minimize aggregate segregation. Specifications are often provided to aggregate manufacturers and contractors mentioning the desired storage and stockpiling practices. For example, the New Hampshire Department of Transportation standard specifications on base courses (http://www.nh.gov/ dot/org/projectdevelopment/highwaydesign/specifications/ documents/2010_Division_300.pdf) include the following requirements for aggregate stockpiling: Stockpiles shall be constructed in layers that minimize segrega- tion. The desired optimum thickness of layers is 6 ft. (1.8 m) and in no instance shall the layer be more than 10 ft. (3 m). Each layer shall be completed before the next layer is started. Construction of stockpiles by direct use of a fixed conveyor belt system or by dumping over a bank will not be permitted. Similarly, stockpiling practices recommended by the Ala- bama Department of Transportation (http://www.dot.state. al.us/mtweb/Testing/testing_manual/doc/pro/ALDOT175. pdf) include • Stockpiles need to be placed on firm, well-drained ground that is free of any material that could cause contamination. • Stockpiles should be built in layers of uniform thick- ness and not in cone-shaped piles that result in segrega- tion of piles. • After the first layer of the stockpile is placed, it is impor- tant that heavy transporting equipment not be allowed to run on top of this layer because this tends to degrade the aggregate by grinding the particles together, also con- taminating the aggregate with mud and other deleterious substances from the wheels or tracks of the vehicle. • If the stockpile is to be constructed in more than one layer in height, the aggregate should be dumped in a small pile at the base of the stockpile and then moved over to the stockpiled layer in place by a crane equipped with a clam- shell, front-end loader or bulldozer equipped with large pneumatic tires. The standard operating procedures recommended by the GDOT (http://www.dot.ga.gov/doingbusiness/Materials/ Documents/StudyGuide9_22_04.pdf) provide graphical rep- resentations of the prohibited and recommended practices as far as aggregate stockpiling and aggregate sampling from different types of stockpiles are concerned (see Figure 26). Construction Practices as a Source of Segregation Different construction practices can contribute significantly to aggregate segregation and therefore should be controlled through the agency specifications. White et al. (2004) observed that aggregate trimming operations are likely to contribute the most to the segregation problem as they shake the aggregate, causing fine particles to migrate to the bottom of the layer. Sub- sequent removal of the top aggregate by the trimmer leaves the fine aggregate behind, resulting in uneven spatial distribution FIGURE 24 Spatial distribution of particle gradation in a stockpile (modified from Nohl and Domnick 2000).

FIGURE 25 (a) Windrow configuration and (b) segregation resolution in a windrow pile (Nohl and Domnick 2000). (a) (b) FIGURE 26 Example specifications regarding aggregate stockpiling (Georgia DOT).

43 of aggregate particle sizes in the constructed layer. They also observed that low moisture in the aggregate mostly corre- sponded to increased segregation as a result of poor adhesion between finer and larger particles. Through in situ testing of full-scale unbound aggregate test sections, they suggested changes to construction operations to limit spatial variations in constructed layer properties. These changes include (1) limiting movement of aggregate by pri- marily transporting aggregate transversely, rather than longi- tudinally, and (2) moistening the aggregate before trimming to reduce fines migration. Williamson and Yoder (1967) studied the achieved compaction levels in different rigid pavement subbase layer constructions in the state of Indiana and con- cluded that the lack of compaction could be attributed to non- uniform aggregate gradations in the constructed layer, which is an indicator of segregation during the construction process. Aggregate Degradation and Possible Sources Aggregate degradation is defined as the breakdown of an aggregate into smaller particles (Barksdale 1991). Aggregate degradation can occur during the process of aggregate stock- piling or during the placement and compaction of aggregate base and subbase layers. Formation of stockpiles by pushing of aggregates using dozers and handling of the aggregates during different stages of construction both may result in degradation of the larger particles into smaller fractions. Although the prob- lem of degradation is not as severe for quarries excavating hard rock formations, the problem can be significant for operations dealing with “softer” parent rocks. Some quarries compensate for the potential degradation by producing aggregate sizes that are coarser than the target aggregate size. Moreover, different agencies impose different restrictions on the type of construc- tion vehicle allowed to operate on aggregate stockpiles. Compaction of constructed layers may impose heavy loads that cause aggregate degradation. Aughenbaugh et al. (1963) indicated that degradation is dominant in the top lift of an aggregate layer. Thus, the height of a layer during compaction may contribute toward nonuniform aggregate gradation result- ing from degradation of individual particles. This is particularly critical for aggregate layers constructed with large lift thick- nesses. To ensure adequate compaction at greater depths, high compactive energies need to be imparted on the layer surface. Such high compactive energies result in significant crushing of aggregate particles near the layer surface, changing the gra- dation and thus achieved density. Density-based compaction control techniques may give erroneous indications of layer compaction in such cases because density measuring devices, such as the nuclear density gauge, can measure the compac- tion level for only the upper few inches (typically 12 in. for a nuclear density gauge) and do not check the compaction levels for deeper sections in the layer. Thus, it is important to control the amount of energy imparted to the aggregate layer during the compaction process through adjusting the amplitude and frequency of impacts applied by vibratory rollers. CONSTRUCTION LIFT THICKNESS AND ITS EFFECT ON COMPACTABILITY Background One of the primary factors affecting the performance of UAB and subbase layers is the DOC. Aggregate materials received from the source are placed on the prepared subgrade and com- pacted to the design layer thicknesses. Generally, upon compac- tion an unbound aggregate layer loses approximately one-third of its loose placement depth (Barksdale 1991; Saunders 1997). Specifications require the compaction to be carried out imme- diately after placement of the aggregate material while the gradation and moisture content are still at the specified values (NSSGA 1989). Maximum allowable lift thicknesses are usu- ally specified during the construction of unbound aggregate layers to ensure adequate compaction, which is critical to pavement layer performance under loading. Saunders (1997) reports that the maximum lift thicknesses specified by agen- cies most likely were established in the early days of highway construction, when only static rollers or limited vibratory roll- ers were available. Saunders also indicates that in view of the modern construction equipment, these maximum lift thickness thresholds most likely are on the conservative side. Wells and Adams (1997) successfully constructed aggregate base courses in single-lift depths of 10 and 12 in., which were greater than the 8-in. maximum allowed by North Carolina Department of Transportation (NCDOT) specifications. Similarly, Womack (1997) reports successful construction of aggregate base courses in Virginia with 10-in. thick lifts, which were greater than the maximum construction lift thicknesses specified by the Virginia Department of Transportation (VDOT) at that time. Researchers have since focused on evaluating the effective- ness of layer compaction when aggregate layers are constructed with large lift thicknesses. Bueno et al. (1998) constructed test pads in Texas and Georgia using crushed limestone and crushed granite, respectively, with lift thicknesses ranging from 6 to 21 in. Three test strips were constructed in Georgia with differ- ent lift thicknesses. A “target test strip” was compacted in two lifts: one 178-mm (7-in.) compacted lift under a 152-mm (6-in.) compacted lift. Two other test sections (sections 1 and 2) were both compacted in one single lift to a final compacted thick- ness of 330 mm (13 in.). In-place density measurements indi- cated compaction levels of 102.5%, 103.9%, and 103.3% for the target Strip, Test Section 1, and Test Section 2, respectively, Key Lessons • Aggregate segregation and deterioration can be minimized through proper stockpiling and construc- tion practices. • Stockpiling of aggregates using the windrow concept has been proven to be the most efficient practice as far as minimizing segregation is concerned.

44 indicating that adequate compaction levels could be achieved even for higher construction lift thicknesses. Constructing test pads in Texas, Bueno et al. (1998 and 1999) observed that higher densities could sometimes be achieved for 457-mm (18-in.) and 584-mm (23-in.) lifts compared with those achieved for 305-mm (12-in.) lifts. In general, dry densities were found to increase with depth into the compacted base, illustrating that the lower parts of the aggregate base course were being compacted even for greater lift thicknesses. This trend was supported by the shear wave velocity profiles obtained from Spectral analysis of surface waves (SASW) testing. From SASW data, they observed that stiffness of a graded aggregate base (GAB) course was sensi- tive to moisture variations and concluded this sensitivity likely was caused by changes of effective confining stress, which occur when the material is wetted or dried. Overall, findings from this study clearly indicate that thicker single lifts could be compacted equally well or sometimes better than are the thinner aggregate layers commonly constructed by agencies. Allen et al. (1998) conducted a survey of all state trans- portation agencies and found that 12 of 36 responding states allowed a maximum lift thickness of 6 in. or less, one state allowed 7-in. lifts, and 16 states allowed 8-in. lifts. Only three states allowed thicker lifts (Washington, 9-in.; North Carolina, 10-in.; and Maine, 12-in.). The survey of state and Canadian provincial agencies conducted under the scope of the cur- rent synthesis study found that 11 of 46 responding agencies allowed construction lift thicknesses of 8 in. The current lim- its for construction lift thickness in North Carolina and Maine were the same as those reported by Allen et al. (1998). Note that as of 2012, 17 agencies still restricted the maximum lift thick- ness to 6 in. Figure 27 summarizes all the data collected from the survey respondent states and Canadian provincial agencies. From the figure it is evident that despite several research and trial studies demonstrating the effectiveness of aggregate layer construction with larger lift thicknesses, the current practices in state and Canadian provincial transportation agencies still use a conservative approach in this regard. Thus, more research and demonstration projects need to focus on the advantages and disadvantages (if any) of higher construction lift thicknesses. Such studies will also help harmonize the construction prac- tices throughout the United States and Canada. From successful implementation of thick single-lift aggre- gate base and subbase layer construction, Allen et al. (1998) recommended the following changes to unbound aggregate layer construction practices: • Equipment: Mixing is to be accomplished by stationary plant, such as a pugmill, or by road mixing using a pug- mill or rotary mixer. Mechanical spreaders should be used to avoid segregation and achieve grade control. Suitable vibratory compaction equipment should be employed. • Mixing and Transporting: The aggregates and water should be plant mixed (stationary or roadway) to the range of optimum moisture plus 1% or minus 2% and transported to the job site so as to avoid segregation and loss of moisture. • Spreading: The material is to be placed at the specified moisture content to the required thickness and cross sec- tion by an approved mechanical spreader. At the engi- neer’s discretion, the contractor may choose to construct a 500-ft long test section to demonstrate achieving ade- quate compaction without particle degradation for lift thicknesses in excess of 13 in. The engineer may allow thicker lifts on the basis of the test section results. Optimum Construction Lift Thickness As observed from the survey results, no consensus exists among transportation agencies with regard to the maximum allowed construction lift thickness for UAB/subbase layers. From exten- sive review of literature conducted under the scope of this syn- thesis study, it was observed that most research studies and trial implementation projects could successfully compact 12-in. thick aggregate layers while achieving desired compaction levels. Thus, it is suggested that 12-in. aggregate lifts be standardized as “optimum construction practice” for UAB/subbase layers. Given adequate support conditions, construction of unbound aggregate layers in such thick lifts could sufficiently expedite 37% 24% 9% 13% 15% 2% (17) (11) (4) (6) (6) (1) 0 10 20 30 40 0% 20% 40% 60% 80% 100% 6 in. 8 in. 10 in. 12 in. Other (please indicate) No such restrictions Number of Responses Percentage of Respondents 46 survey respondents FIGURE 27 Maximum construction lift thickness allowed for unbound aggregate layers.

45 the construction process while ensuring adequate compaction levels. For aggregate layers being constructed over “firm” pre- pared subgrades (often represented by subgrade CBR > 8%), the compaction of 12-in. thick layers may be possible. However, reduced lift thicknesses may need to be adopted for “weaker” subgrade support conditions (subgrade CBR < 8%). DOCUMENTED AGGREGATE BASE AND SUBBASE LAYER CONSTRUCTION PRACTICES Figures 28 to 30 summarize agency responses to various cur- rently adopted aggregate base and subbase layer construction practices. Figure 28 highlights that 52.2% of the respon- dent agencies allow construction of two functionally differ- ent aggregate layers on top of each other, such as an OGDL underlain by a dense-graded aggregate subbase, in pavements. Furthermore, 66.7% of the respondent agencies do not sepa- rate two unbound aggregate layers by any kind of constructed aggregate separation or filter layers (see Figure 29). Figure 30 indicates that 65.2% of the respondent agencies allow construc- tion of unbound aggregate layers directly over or under pave- ment layers stabilized or treated with lime, fly ash, cement, or bitumen. Note that these findings have implications on some of the domestic and foreign innovative pavement construction practices, such as the construction of “inverted pavements.” INVERTED PAVEMENTS Sustainable application of unbound aggregate structural lay- ers in pavements would improve the designs of low, medium, and moderately high volume roads while reducing the depen- dence on asphalt (and thus crude oil) for pavement construc- tion. Such an alternative is offered by an inverted pavement section that consists of an unstabilized crushed stone base or GAB sandwiched between a lower cement-stabilized layer and a thin upper asphalt concrete surfacing. Conceptual Background Conventional pavement systems rely on a combination of asphalt concrete or PCC and aggregate base components to transfer load to the subgrade. All three components use Key Lessons • No common practice exists among transportation agencies as far as the maximum construction lift thick- ness of UAB/subbase layers is concerned. • Maximum construction lift thickness value for UAB/ subbase layers are best based on project “test-strip” sections using the specific materials and equipment. • From extensive review of literature and state prac- tices, this synthesis study suggests an optimum construction lift thickness of 12 in. for UAB/subbase layers. Note that this suggestion is based on the assumption that the UAB/subbase layer to be con- structed is at least 12-in. thick. Moreover, the DOC achieved is contingent upon the use of adequate equipment by the contractor. 52% (24) 26% 22% (12) (10) 0 10 20 30 40 0% 20% 40% 60% 80% 100% Yes No Other Number of Responses Percentage of Respondents 46 survey respondents FIGURE 28 Agency responses to whether multiple unbound aggregate layers are allowed to be placed on top of each other. 17% (4) 67% (16) 17% (4) 0 10 20 0% 20% 40% 60% 80% 100% Yes No Other Number of Responses Percentage of Respondents 24 survey respondents FIGURE 29 Agency responses to whether the two unbound aggregate layers are separated by any kind of constructed aggregate separation/filter layers.

46 aggregates as their primary constituent. Classic pavement design places higher modulus, more durable layers toward the surface. Inverted pavement is a composite system composed of asphalt top layer(s) and a well-compacted unbound crushed stone base layer over a stiffer bound subbase that is usually cement treated. Mechanistically, this configuration provides a stronger reaction platform than unbound subgrades or sub- bases, allowing increased granular base compaction during construction, and it also has the potential to take advantage of the compressive stresses induced in the granular aggregate base owing to the presence of the stiff underlying layer. This pavement design philosophy potentially offers economic advantages by requiring less asphalt concrete and placing the burden of strength and structural performance on relatively less expensive underlying layers. Inverted pavements were first introduced in South Africa and involved the construction of thick crushed stone base lay- ers over stabilized subbase layers. The superior performing crushed stone base layers used in South African inverted pave- ments are also known as “G1” base layers (Horne et al. 1997; Jooste and Sampson 2005; De Beer 2012). These pavements are also called stone interlayer pavements, G1-base pavements, inverted base pavements, sandwich pavements, and upside down pavements (Lewis et al. 2012). Figure 31 shows the layer configuration of a typical inverted pavement structure. As shown in Figure 31, the HMA layers in inverted pave- ment sections often are very thin, so their contribution to the structural capacity of the pavements often is not significant. Primarily, these surface layers provide a smooth ride qual- ity and protect the underlying pavement layers from water infiltration. The unbound aggregate layer is the primary load- bearing layer in inverted pavement structures. Summarizing the construction practices and layer configurations of these pavement systems, De Beer (2012) presented the following definition for inverted pavements: A structural pavement system, where the static modulus of the unbound base layer is lower compared with the supporting (mainly lightly cementitious) subbase layers. Unbound base layer (crushed rock) of extremely high bearing capacity is usually covered with 12 mm to 50 mm asphalt layer for sealing and functional properties. Owing to the reduced thickness of the HMA surface layers, these pavement systems are cost-effective alternatives for high- performance pavement structures. The primary advantages of inverted pavements include (1) better compaction of unsta- bilized materials placed over the stabilized layers; (2) opti- mum use of unstabilized crushed stone; and (3) elimination or significant reduction in reflective cracking in the pavement structures (Barksdale and Todres 1983). Response Mechanism The UAB is primarily a structural load-carrying component in inverted pavement sections. When properly compacted, the UAB causes lateral dissipation of traffic-induced stresses through interparticle contact points. The stiff UAB and the cement-stabilized subbase combined result in a significant reduction in the vertical compressive stress levels on top of the subgrade, thus eliminating chances of pavement fail- ure because of subgrade rutting. However, the UAB in an inverted pavement structure is subjected to considerably higher stresses to make the base layer prone to rutting, a potential failure mechanism for inverted pavement sections. Thus, construction procedures for inverted pavements aim to eliminate rut accumulation within the UAB layers through innovative compaction procedures. The thin HMA surface typically considered in an inverted pavement section induces considerably high stress states within the aggregate base under wheel loading. Owing to the stress-hardening nature of unbound aggregates, these high stress states often lead to the aggregate layer developing high elastic modulus values, often on the order of 689 MPa 65% (30) 22% (10) 13% (6) 0 10 20 30 40 0% 20% 40% 60% 80% 100% Yes No Other Number of Responses Percentage of Respondents 46 survey respondents FIGURE 30 Agency responses to whether the construction of unbound aggregate layers is allowed over or under pavement layers stabilized or treated with lime, fly ash, cement, or bitumen. FIGURE 31 Layer configuration of a typical inverted pavement section.

47 or 100 ksi (Maree et al. 1982a, 1982b; O’Neil et al. 1992). Such high modulus levels achieved in the aggregate bases of inverted pavement sections would help better dissipate the traffic-induced stresses with depth. Moreover, the pres- ence of the stiffer subbase layer in an inverted pavement section causes the neutral axis in bending to fall below the aggregate base layer. This results in the surface layers and the UAB layers performing mainly under compres- sion. Accordingly, the stiffness profile in inverted pave- ment structures prevents the development of tensile stresses in the UAB even if a linear model is used to represent it (Cortes 2010). The stiff aggregate base layer also leads to a reduction in the tensile stresses at the aggregate base course-HMA surface interface, thus significantly reducing the chances for reflective cracking occurring in these pave- ment structures. Material Specifications and Construction Procedure Material Specifications The aggregate material to be used in the base course of an inverted pavement structure is obtained from crushing of hard, sound, durable, and unweathered parent rock. All the faces of the aggregate particles are required to be fractured. South African G1 base specifications allow the material gra- dation to be adjusted only through the addition of fines pro- duced from the crushing of the original parent rock. Table 3 lists the (a) particle size distribution and (b) other material quality specifications used in South Africa for use in G1 base course applications (TRH 1985; Buchanan 2010). Note that the gradation requirements listed in Table 3a are based on restricting the “n” values in the Fuller’s or Talbot’s equation (as defined in Equation 1) between 0.33 and 0.50. Note that in Equation 1, P is the percentage (%) of material by weight finer than the sieve size being con- sidered; d is the sieve size being considered; D is the maxi- mum aggregate particle size in the current matrix; and n is a parameter that adjusts the gradation curve for fineness or coarseness. P dD n( )= × 100 (1) Construction Procedure Compaction of the UAB layer in an inverted pavement struc- ture is the most critical step during its construction to ensure that individual layers perform as desired. The UAB is con- structed on top of a stabilized subbase, which provides a solid construction platform for the placement and compac- tion of the UAB layer and ensures that adequate density lev- els can be achieved. The DOC achieved in the UAB layer is dependent on the energy applied, as well as the initial and final gradations of the aggregate material used. (a) Sieve Size (mm) Sieve Size (in.) Percent Passing G1, 37.5 NMS G1, 26.5 NMS 50 2.0 37.5 1.5 100 26.5 ~1.0 84–94 100 19.0 ¾ 71–84 85–95 13.2 ~1/2 59–75 74–84 4.75 #4 36–53 42–60 2.00 #10 23–40 27–45 0.425 #40 11–24 13–27 0.075 #200 4–12 5–12 (b) Sources: TRH 1985; Buchanan 2010. Aggregate Material Property Specified Threshold Values Minimum 10% FACTa 110 Maximum aggregate crushing valueb 29% Liquid limit <25 Linear shrinkagec <2% Plasticity index (PI) <4 a 10% FACT (fines aggregate crushing value) is the force in kilonewtons required to crush a sample of aggregate passing the 13.2-mm and retained on the 9.5-mm sieve so that 10% of the total test sample will pass a 2.36-mm sieve. b The aggregate crushing value (ACV) of an aggregate is the mass of material, expressed as a percentage of the test sample that is crushed finer than a 2.36-mm sieve when a sample of aggregate passing the 13.2-mm and retained on the 9.50-mm sieve is subjected to crushing under a gradually applied compressive load of 400 kN. c The linear shrinkage of a soil for the moisture content equivalent to the liquid limit is the decrease in one dimension, expressed as a percentage of the original dimension of the soil mass, when the moisture content is reduced from the liquid limit to an oven-dry state. TABLE 3 RECOMMENDED PARTICLE SIzE DISTRIBUTION RANGE FOR SOUTH AFRICAN G1 BASE

48 The compaction of unbound aggregate layers in the South African inverted pavement structures involves the following two phases: standard compaction phase and particle inter- locking or slushing phase. The standard compaction phase is carried out using a combination of grid rollers, vibratory roll- ers, and pneumatic tire rollers. Commonly two to three passes of the grid roller are used to gently knead the aggregate layer into shape. Subsequently, the vibratory roller is used to com- pact the layer to 85% of apparent solid density. It is important to note that the amplitude and frequency of vibration need to be strictly monitored during this phase because too much vibration can easily lead to “de-densification” of the aggre- gate matrix. Moreover, extreme care is to be exercised to pre- vent the breakage of individual aggregate particles under the vibratory roller. The aggregate moisture content usually is maintained near the “optimum” conditions during this phase of compaction to aid the rearrangement of individual aggre- gate particles into a densely packed matrix. The fines fraction in the aggregate matrix plays a critical role during this phase by lubricating the aggregate contact points. Thus, it is impor- tant that the aggregate material used in the construction of these superior performing base layers contain the adequate amount of fines. A rule of thumb used in the construction of South African G1 base course layers is that OMC values less than 4% are indicative of too few fines in the aggregate matrix, whereas OMC values higher than 6% are indicative of too many fines (De Beer 2012). The second phase of compaction involves consolidating the material under saturated conditions by expelling or “slush- ing” out the excess fines material from the matrix, allowing the larger particles to interlock into a “superdense” matrix. The fines serve as lubricants to ensure reorientation and interlocking of the larger particles into a superdense matrix. This “washing out” of the fines, accompanied by compac- tion, is continued until the water draining from the pavement becomes colorless and does not contain any trace of excess fines (De Beer 2012). A pneumatic tire roller passing over the aggregate layer without leaving any indentations is used as an indicator of the achievement of adequate compaction. South African specifications require achieved density to be greater than 88% of solid particle density (assuming solid rock for the equal volume with no voids). It is usual to place the prime coat and HMA surfacing layer immediately after compaction of the UAB layer. This is primarily because the aggregate base layer is noncohe- sive in nature, and the aggregate matrix may get disturbed upon exposure to direct application of traffic loads and weathering. Previous Findings on the Benefits of Inverted Pavements The first application of inverted pavements in the United States can be traced back to 1954 in New Mexico (Barksdale and Todres 1983). These initial inverted pavement sections involved the overlaying of several badly broken concrete pavements with 152 mm (6 in.) of unstabilized granular base, and 51 mm (2 in.) of asphalt concrete. Johnson (1960) reported that after six years of heavy traffic, no reflection cracking or significant rutting had developed in the test sections. Subsequently, two experimental roads were constructed in New Mexico in about 1960, consisting of a 76-mm (3-in.) asphalt concrete surfacing, 152-mm (6-in.) granular base, and a 152-mm (6-in.) granular subbase treated with 4% cement. U.S. Army Corps of Engineers Experience The U.S. Corps of Engineers studied the behavior of the various layers in flexible pavement structures having lime- stabilized and cement-stabilized subbases: that is, inverted base type structures (Ahlvin et al. 1971; Barker et al. 1973; Grau 1973). The objective of the study was to measure the mechanical response of full-scale pavement structures and compare the results against predictions from layered elastic theory and other available constitutive models. Two inverted base pavement structures were investigated, both composed of a 90-mm asphalt concrete layer, a 150-mm crushed lime- stone base, a 380-mm stabilized clay subbase, and a clay sub- grade (CBR of 4%). The structures were subjected to traffic under controlled conditions while monitoring displacements and stresses at key locations (Ahlvin et al. 1971; Barker et al. 1973; Grau 1973). Linear elastic analyses failed to adequately predict the measured stresses and strains in different layers and the plastic subgrade deformation. The performance of the inverted pavement structures was found to be influenced by the stiffness and tensile strength of the cement-treated base. This study highlighted the importance of a compre- hensive material characterization and numerical implementa- tion through appropriate constitutive models. Furthermore, it urged the development of laboratory tests capable of simulat- ing field conditions and the introduction of nonlinear models in numerical simulations (Barker et al. 1973). Barksdale and Todres Barksdale and Todres (1983) constructed 12 laboratory-scale instrumented pavement structures and cyclically loaded them to failure under controlled environmental conditions. Among conventional flexible pavement and full-depth asphalt pave- ment test sections, they also tested two inverted pavement test sections made of 89-mm thick asphalt concrete layers over 203-mm thick unbound aggregate layers (well-graded gra- nitic gneiss), over a 150-mm thick cement-stabilized subbase, over a micaceous nonplastic silty sand subgrade. One inverted pavement section had a 152-mm (6-in.) thick cement stabilized crushed stone subbase; the other had a 152-mm (6-in.) thick cement-treated, silty sand subbase. It was found that the cement- treated base facilitated compaction in inverted structures lead- ing to denser unbound aggregate layers ( Barksdale 1984).

49 The pavement sections were subjected to a 28.9-kN cyclic load for the first 2 × 106 repetitions, followed by cyclic appli- cation of a 33.4-kN load until failure. Monitoring the per- formance of the test sections under loading, Barksdale and Todres observed that the two inverted pavement sections outperformed equivalent pavement structures in terms of lower resilient surface displacements, reduced transferred compressive stress onto the subgrade, and less tensile radial strain at the bottom of the asphalt concrete layer (Barksdale and Todres 1983; Avellandeda 2010). The superior mechani- cal performance of the inverted pavement structures was clearly reflected from the significantly higher number of load cycles to failure (3.6 × 106 and 4.4 × 106) compared with the best performing conventional flexible pavement section (Barksdale 1984; Tutumluer and Barksdale 1995). Tutumluer and Barksdale Tutumluer and Barksdale (1995) conducted numerical mod- eling of the two full-scale instrumented inverted pavement sections tested by Barksdale and Todres (1983), and made the following observations: • Cement-stabilized inverted sections can successfully withstand large numbers of heavy loadings through – Lower vertical subgrade stresses owing to the “beam action” of the stiff base layer; – Lower tensile strain at the bottom of the asphalt layer; and – Lower resilient surface deflections. • The upper portion of the cement-treated subbase and almost all of the unstabilized crushed stone base near the load were in horizontal compression. The bottom half of the subbase and a thin layer on top of the sub- grade were in horizontal tension. • Presence of the cement-stabilized layer beneath the aggregate base resulted in horizontal compressive stresses of magnitudes ranging from 0 to 110 kPa (0 to 16 psi) in the unstabilized crushed stone base. This was probably a major factor contributing to the lower permanent deformation and higher resilient moduli of these base layers as observed from laboratory testing. • From sensitivity analyses conducted using the GT- PAVE finite element (FE) program, they observed that the optimum and economical inverted pavement sec- tion constructed over a weak subgrade would consist of an unstabilized aggregate base 152 mm (6 in.) thick and a 152-mm to 203-mm (6-in. to 8-in.) thick cement- stabilized subbase. Lafarge Quarry Access Road —Morgan County, Georgia Two 122-m (400-ft.) long inverted pavement test sections were constructed on a new access road at the Lafarge Build- ing Materials quarry near Madison, Georgia, in 2001. Both test sections had a 200-mm (8-in.) thick cement-treated base layer, a 150-mm (6-in.) thick GAB layer, and a 75-mm (3-in.) thick HMA layer. The only difference between the two inverted pavement sections was in the construction of the GAB layer: the first section was constructed using the South African “slushing” technique, whereas the second section was constructed using conventional construction methods. Terrell et al. (2003a, b) conducted miniaturized versions of traditional cross-hole and downhole seismic tests to deter- mine the stiffnesses of each base layer. Horizontally propagat- ing compression and shear waves were measured under four different loading conditions to determine Young’s moduli and Poisson’s ratios of the base. An increase in stiffness with an increase in load was measured. In addition, it was found that the Georgia and South Africa sections had similar stiffnesses. Surprisingly, the traditional section was found to be some- what stiffer than the other sections. This higher stiffness was thought to be caused by a prolonged period of compaction before construction of the UAB layer, which essentially trans- forms the traditional section (Terrell et al. 2003b). Comparing the performances of the two inverted pave- ment test sections with a conventional flexible pavement sec- tion subjected to the same loading, Lewis et al. (2012) made the following observations: • The two test sections performed remarkably well for more than 10 years, without needing any maintenance or resurfacing; • No significant rut accumulation was observed in the inverted pavement test sections, whereas the conven- tional pavement section exhibited both “minor” and “major” rutting problems; • FWD testing conducted in 2007 indicated that the two inverted pavement sections had remaining service lives of 99.34% (conventional compaction) and 94.61% (compacted using the South African slushing method), respectively, whereas the conventional pavement sec- tion had a remaining service life of 67.92%. FHWA International Scanning Tour A scanning study of France, South Africa, and Australia spon- sored by FHWA, AASHTO, and NCHRP investigated innova- tive programs for pavement preservation (Beatty et al. 2002). During the scanning tour, the team observed typical pave- ment structures used by the countries visited to ensure longer- lasting, better-performing pavement systems. Figures 32 and 33 show the typical pavement structures constructed by Australia and South Africa, respectively, as noted by Beatty et al. (2002). As can be seen from the figures, it is common practice in Australia and South Africa to use thick aggregate layers in conjunction with relatively thin HMA surface lay- ers. The practice in Australia involves the use of multiple

50 unbound aggregate layers in conjunction with a thin HMA surface layer, whereas the South African practice involves the construction of inverted pavement sections. Application of Stone Interlayer Pavements in Louisiana Stone interlayer pavement designs were introduced in Loui- siana to reduce the problem of reflective cracking that is often observed in flexible pavements constructed using soil- cement bases (Rasoulian et al. 2000, 2001). Titi et al. (2003) compared the performances of stone interlayer pavements (3.5-in. HMA surface layer; 4-in. crushed limestone inter- layer; 6-in. in-place cement-stabilized base course layer; and 12-in. lime-treated subgrade layer) with conventional flexible pavements with cement-treated bases (3.5-in. HMA surface layer, 8.5-in. in-place cement stabilized base course layer, and 12-in. lime-treated subgrade layer) constructed on State Highway LA-97 near Jennings, Louisiana. Both pavements were monitored for more than 10 years and were evaluated through pavement distress surveys, testing for roughness and permanent deformation, as well as evaluation of pavement structural capacity through dynamic nondestructive testing (NDT). The same two designs were also compared through accelerated pavement testing at the Louisiana Transportation Research Center. Through analyses of the field monitoring and accelerated testing data, Titi et al. (2003) reported that the stone interlayer pavements performed significantly bet- ter than did the conventional pavement designs with cement- treated base. From comparing the performances of the two pavement types, Titi et al. (2003) made the following pri- mary observations: • Both pavement types showed an increasing trend in crack accumulation with time. However, the rate of crack accu- mulation was significantly lower for the stone interlayer pavement sections. • The average International Roughness Index value for the stone interlayer pavement was lower than that for the con- ventional flexible pavement after 10.2 years. This indi- cated smooth surface conditions and better ride quality for the stone interlayer pavement. This was attributed to the lower amount of reflective cracking in the stone inter- layer pavement. • The stone interlayer pavement could withstand about four times the number of load applications (1,294,800 ESALs) under accelerated testing compared with the conven- tional flexible pavement section (314,500 ESALs) before undergoing failure. • From survival analyses of the two accelerated pave- ment sections, Metcalf et al. (1998) concluded that the dominant mode of failure (88%) for the stone interlayer pavement was rutting, whereas that for the conven- tional pavement was cracking. • Through regression analyses of the long-term perfor- mance data of the two test sections along LA-97, it was concluded that the only mode that could lead to the fail- ure of both of the modes was cracking. The regression analyses clearly established the superior performance of stone interlayer pavement sections. FIGURE 32 Typical heavy-duty pavement configuration in Australia (Beatty et al. 2002). FIGURE 33 Typical pavement sections in South Africa constructed with high-quality crushed aggregate base layers (Beatty et al. 2002).

51 • The initial material cost for the stone interlay pavement was approximately 20% higher than that for the conven- tional pavement. However, considering the significantly higher number (300% higher) of load applications until failure, the stone interlayer pavement alternative indi- cated considerable savings when life-cycle costs were analyzed. LaGrange Bypass Project, Troup County, Georgia Encouraged by the positive results from the inverted pavement test sections in Morgan County, Georgia, in 2009 GDOT con- structed another inverted pavement test section on the South LaGrange Loop in Troup County. The constructed inverted pavement test sections had (1) 150-mm (6-in.) thick stabilized subgrade; (2) 250-mm (10-in.) thick cement-treated base; (3) 150-mm (6-in.) thick GAB; (4) 50-mm (2-in.) thick Super- pave binder course; and (5) 37-mm (1.5-in.) thick Superpave surface course (Lewis et al. 2012). The GAB was constructed using standard construction techniques at a moisture content of 100% to 120% of the OMC. Figure 34 shows a schematic of the inverted pavement sections constructed as part of this project. Buchanan (2010) compared the life-cycle costs for the LaGrange Bypass inverted pavement sections with a rigid pavement designed to carry the same amount of traffic (the rigid pavement had a 9.5-in. thick PCC slab over a 10-in. thick GAB over a 6-in. thick prepared subgrade with a minimum soil support value of 5). Table 4 lists the comparative cost estimates over a 30-year life cycle as presented by Buchanan (2010). As can be seen from the table, the inverted pavement section results in net savings of $139,000 over a 30-year period. Avellandeda (2010) developed new field test methods to characterize the stress-dependent stiffness of UAB layers in these inverted pavement test sections and found that inverted pavement sections could exceed the structural capacities of flexible pavement designs and result in savings to 40% of the initial construction costs. Lewis et al. (2012) reported that the test sections showed excellent structural capacities and long remaining lives upon FWD testing immediately after construction. Cortes (2010) conducted precompaction and postcompaction sieve analyses of aggregate samples collected from the GAB and reported inconclusive data about the extent of particle crushing. By digging trenches through the HMA layers to expose the GAB and subsequently processing the grain skeleton photographs through digital image analysis, Cortes found evidence of compaction-induced anisotropy in the GAB as the coarse aggregate particles were found to preferentially align their major axis parallel to the horizontal plane. Through FE analy- ses of the test sections, Cortes observed that both vertical and radial stresses in the UAB layer remained in compression throughout the layer depth. Luck Stone Bull Run Project, Virginia An application in 2010 of inverted pavement in Virginia involved a relocated road (Virginia Highway 659) bypassing the Luck Stone Bull Run Quarry. The project included collab- oration between Luck Stone, Texas A&M University (ICAR), FHWA Office of Infrastructure R&D, Virginia DOT, and the Virginia Transportation Research Council. This section was designed using the ICAR model, and the materials charac- terization protocol was carried out at Texas A&M Univer- sity and the Texas Transportation Institute and instrumented heavily through FHWA sponsorship. Discussing the benefits of this inverted pavement trial application, Weingart (2012) reported a potential for 22.3% cost savings compared with the construction of a conventional flexible pavement with equiv- alent structural and functional capacities; the estimated cost for construction of the conventional flexible pavement section was $21,311 per 100 linear ft, whereas that for the inverted pavement section was $16,555 per 100 linear ft. FIGURE 34 Schematic of inverted pavement section constructed in LaGrange Bypass Project, Troup County, Georgia. Event Cost ($/Lane-Mile) Inverted Pavement PCC Pavement Installation cost 342,000 584,000 10 years of maintenance 101,000 20 years of maintenance 123,000 20–30 years of maintenance 121,000 30-year life-cycle cost 566,000 705,000 Net savings 139,000 TABLE 4 LIFE-CYCLE COST COMPARISON FOR LAGRANGE BYPASS INVERTED PAVEMENT SECTION WITH A RIGID PAVEMENT SECTION DESIGNED TO SUSTAIN THE SAME TRAFFIC LEVEL OVER A 30-YEAR PERIOD

52 Summary of Past Experience on Inverted Pavements From extensive review of literature covering inverted pave- ment applications internationally as well as within the United States, it was observed that almost all applications of inverted pavements have resulted in favorable performance compared with conventional pavement structures. In addi- tion to resulting in superior performance, inverted pavement sections often have led to significant cost savings over the life cycle of the pavement. Although the construction of pave- ments using thick unbound aggregate layers appears to be a common practice in countries such as Australia and South Africa, projects involving such pavements have been con- fined primarily to trial studies in the United States. Moreover, these trial projects have been confined to a limited number of states, with most other states showing resistance to the adop- tion of such innovative pavement construction practices. Possible explanations of why inverted pavements have not been constructed in the United States include (1) traditional pavement designs and construction practices using rather thick asphalt or concrete surface courses were still afford- able; (2) details of foreign technology related to UAB com- paction, such as the South African slushing technique, were not readily available; and (3) cement-treated subbase used in inverted pavements was considered a potential risk for pavement cracking, especially in northern climates. A con- scious effort needs to be made to encourage the construction of inverted pavements in the United States to fully study any potential disadvantages, such as pavement distresses occur- ring as a result of cracking of the cement-treated subbase, perhaps as a result of shrinkage and exposure to freeze-thaw conditions. In addition, in colder climates further evaluation of related pavement design considerations is needed. Accord- ing to the Portland Cement Association, it is possible to limit the percent cement used in inverted subbases (such as to 2% to 3%) to potentially mitigate these cracking problems. The successful construction, ongoing documentation, and tech- nology transfer of the superior performances of the inverted pavement trials no doubt will have a positive impact on such sustainable alternatives to pavement design. Current State of Practice on Alternative Base Course Construction One of the objectives of the current synthesis study was to gather information on the state of practice in the United States and Canada regarding the application of alternative UAB/ subbase layers, such as inverted pavement sections. Accord- ingly, the survey of state and Canadian provincial transpor- tation agencies included questions regarding construction practices such as the South African slushing technique. Only two states (New Mexico and Rhode Island) reported the use of alternative construction techniques. New Mexico DOT reported an ongoing project involving the construction of inverted pavement sections that will use the South African slushing technique for compaction of the UAB layer. Rhode Island DOT indicated that the agency used the “test strip” method to determine the maximum achievable density value for an UAB/subbase layer through repeated compaction of the same spot until no noticeable increase in the density was achieved. The DOC achieved during construction of UAB/subbase layers was then compared with the maximum density values obtained from the test strips. No other state reported the use of innovative construction practices. SUMMARY This chapter presents an overview of current practices as far as material handling and construction practices for UAB and subbase layers are concerned. Extensive review of published literature was conducted to gather information on differ- ent procedures and practices identified by researchers to be adequate/inadequate for pavement layer construction. Aggregate segregation and degradation are identified as two major concerns affecting aggregate gradation, and different practices that magnify these problems are listed. A survey of state and Canadian provincial transportation agencies indicated that only 37% of the responding agencies (46 total respondents) currently have specific guidelines governing aggregate storage, transportation, and stockpiling practices. The current state of the practice regarding construction lift thicknesses indicates a significant gap between the knowl- edge gained through research and trial projects and current agency specifications. Different research studies establishing the effectiveness of greater lift thicknesses during construc- tion are summarized in this chapter and the need for har- monizing such practices among transportation agencies was established. Key Lessons • Inverted pavements involve the construction of a “high quality” crushed stone base layer over a stabi- lized subbase course. • With the aggregate base layer functioning as the primary structural component, inverted pavements offer a long-lasting, economical alternative to con- ventional pavement construction. • Construction of inverted pavements and similar pave- ments utilizing thick crushed aggregate base layers is a common practice in countries such as South Africa, Australia, and France. • All inverted pavement applications in the United States have resulted in equal or better performance compared with equivalent conventional pavement sections. • A conscious effort is required in the United States and Canada to encourage the construction of alterna- tive pavement structures with thick UAB layers as the primary structural component.

53 Finally, this chapter discusses the concept of inverted pavements as an alternative application of UAB layers. The concept behind this application was described, as were the response mechanism and construction procedures. Different research studies emphasizing the effectiveness of inverted pavements are highlighted, and the need for further explo- ration in this area was established. The next chapter dis- cusses the different methods used for the characterization of unbound aggregate materials and layer design. REFERENCES Ahlvin, R.G., W.J. Turnbull, J.P. Sale and A.A. Maxwell, Multiple-Wheel Heavy Gear Load Pavement Tests, U.S. Army Corps of Engineers, Vicksburg, Miss., 1971, 216 pp. Allen, J.J., J.L. Bueno, M.E. Kalinski, M.L. Myers, and K.H. Stokoe II, Increased Single-Lift Thicknesses for Unbound Aggregate Base Courses, ICAR Report No. 501-5, Inter- national Center for Aggregate Research, The University of Texas at Austin, 1998. Aughenbaugh, N.B., R.B. Johnson, and E.J. Yoder, Degrada- tion of Base Course Aggregates during Compaction, School of Civil Engineering, Purdue University, West Lafayette, Ind., 1963. Avellandeda, D.D.C., Inverted Base Pavement Structures, Ph.D. Thesis, School of Civil and Environmental Engi- neering, Georgia Institute of Technology, Atlanta, 2010. Barker, W.R., W.N. Brabston and F.C. Townsend, An Inves- tigation of the Structural Properties of Stabilized Layers in Flexible Pavement Systems, U.S. Army Corps of Engi- neers, 165, Vicksburg, Miss., 1973. Barksdale, R.D. and H.A. Todres, A Study of Factors Affecting Crushed Stone Base Performance, School of Civil Engi- neering, Georgia Institute of Technology, Atlanta, 1983. Barksdale, R.D., “Performance of Crushed Stone Base Courses,” Transportation Research Record 954, Trans- portation Research Board, National Research Council, Washington, D.C., 1984, pp. 78–87. Barksdale, R.D., The Aggregate Handbook, National Stone Association, Washington, D.C., 1991. Beatty, T.L., et al., Pavement Preservation Technology in France, South Africa, and Australia, Office of International Programs, Federal Highway Administration, U.S. Depart- ment of Transportation, and the American Association of State Highway and Transportation Officials, Alexandria, Va., 2002. Buchanan, S., Inverted Pavements-What, Why, and How? AFTRE Industry Education Webinar, Aggregates Foun- dation for Technology, Research, and Education, Alexan- dria, Va., June 1, 2010. Bueno, J.L., K.H. Stokoe, and J.J. Allen, “A Study on the Fea- sibility of Compacting Unbound Graded Aggregate Base Courses in Thicker Lifts that Presently Allowed by State Department of Transportation,” Report No. ICAR 501-2, International Center for Aggregates Research, University of Texas at Austin, 1998. Bueno, J.L., K.H. Stokoe II, J.J. Allen, and M.E. Kalinski, “Eval- uation of Constructing Increased Single-Lift Thicknesses of Unbound Aggregate Bases Case Study in Georgia,” Trans- portation Research Record: Journal of the Transportation Research Board, No. 1673, 1999, pp. 95–102. Cortes, D.D., “Inverted Base Pavement Structures,” Ph.D. Dissertation, School of Civil and Environmental Engi- neering, Georgia Institute of Technology, Atlanta, 2010. De Beer, M., “South African G1 Base Course-Inverted Pave- ment,” 2012 Transportation Research Board Mineral Aggregates Committee (AFP70) Meeting Presentation, 91st Annual Meeting of the Transportation Research Board, Jan. 22–26, 2012, Washington, D.C. Grau, R.W., “Evaluation of Structural Layers of Flexible Pavements,” Miscellaneous Paper, S-73-26, Waterways Experiment Station, 1973. Horne, D., et al., South African Pavement and Other Highway Technologies and Practices, Federal Highway Adminis- tration, U.S. Department of Transportation, Washington, D.C., 1997. Johnson, V.W., “Comparative Studies of Combinations of Treated and Untreated Bases and Subbase for Flexible Pave- ments,” Bulletin 289, Highway Research Board, National Research Council, Washington, D.C., 1960, pp. 44–61. Jooste, F.J., and L. Sampson, “The Economic Benefits of HVS Development Work on G1 Base Pavements,” Department of Public Transport, Roads, and Works, 2005. Lewis, D.E., K. Ledford, T. Georges, and D.M. Jared, “Con- struction and Performance of Inverted Pavements in Geor- gia,” Paper No. 12-1872, Poster Presentation in Session 639, 91st Annual Meeting of the Transportation Research Board, Jan. 22–26, 2012, Washington, D.C. Majidzadeh, K. and C.S. Brahma, Statistical Analysis of Aggregate Size Distribution, The Ohio State University Department of Civil Engineering and Transportation Research Center, 1969. Maree, J.H., N.J.W. van zyl, and C.R. Freeme, “Effective Moduli and Stress Dependence of Pavement Materials as Measured in Some Heavy-Vehicle Simulator Tests,” Transportation Research Record 852, Transportation Research Board, National Research Council, Washington, D.C., 1982a. Maree, J.H., C.R. Freeme, N.J.W. van zyl, and P.F. Savage, “The Permanent Deformation of Pavements with Untreated Crushed-Stone Bases as Measured in Heavy Vehicle Sim- ulator Tests,” Proceedings of the 11th Australian Road Research Board Conference, Melbourne, 1982b. Metcalf, J.M., S. Romanoschi, L. Yongqi, and M. Rasoulian, Construction and Comparison of Louisiana’s Conven- tional and Alternative Base Courses under Accelerated Loading, Interim Rep. 1, Phase 1, Louisiana Transporta- tion Research Center, Baton Rouge, 1998. Miller Warden Associates, “Effects of Different Methods of Stockpile Sampling Interim Report,” Proceedings of Highway Research Board, Highway Research Board, National Research Council, Washington, D.C., 1964.

54 National Stone, Sand and Gravel Association (NSSGA), Stone Base Construction Handbook, NSSGA, Alexandria, Va., 1989. Nohl, J., and B. Domnick, “Stockpile Segregation,” Techni- cal Paper T-551, Superior Industries, Morris, Minn., 2000. O’Neil, D.J., J.P. Mahoney, and N.C. Jackson, An Evaluation of Granular Overlays in Washington State, Final Techni- cal Report, Report No. FHWA-SA-92-042, Washington State Department of Transportation, Olympia, 1992. Rasoulian, M., B. Becnel, and G. Keel, “Stone Interlayer Pavement Design,” Transportation Research Record 1709, Transportation Research Board, National Research Coun- cil, Washington, D.C., 2000, pp. 60–68. Rasoulian, M., H. Titi, M. Martinez, B. Becnel, and G. Keel, Long Term Performance of Stone Interlayer Pavement, Louisiana Transportation Research Center Report 29, Baton Rouge, 2001. Saunders, C.H., “Increasing the Single Lift Thickness for Aggregate Base: Base Placement and Compaction of Increased Lift Thickness of Aggregate Base on Test Proj- ects,” Paper Presented at the 5th Annual Symposium of the International Center for Aggregate Research (ICAR), 1997. Terrell, R.G., B.R. Cox, F.Y.U.H. Menq, J.J. Allen, and K.H. Stokoe, II, “Stiffness of Unbound Aggregate Base Lay- ers in Inverted Flexible Pavements,” International Cen- ter for Aggregates Research 11th Annual Symposium: Aggregates–Asphalt Concrete, Bases and Fines, Austin, Tex, Apr. 27–30, 2003a. Terrell, R.G., B.R. Cox, K.H. Stokoe II, J.J. Allen, and D. Lewis, “Field Evaluation of the Stiffness of Unbound Aggregate Base Layers in Inverted Flexible Pavements,” Transportation Research Record: Journal of the Trans- portation Research Board, No. 1837, Transportation Research Board of the National Academies, Washington, D.C., 2003b, pp. 50–60. Titi, H., M. Rasoulian, M. Martinez, B. Becnel, and G. Keel, “Long-Term Performance of Stone Interlayer Pavement,” Journal of Transportation Engineering, Vol. 129, No. 2, 2003, pp. 118–126. TRH, Guidelines for Road Construction Materials, Draft Technical Recommendation for Highways (TRH) 14, Pre- toria, South Africa, 1985. Tutumluer, E. and R.D. Barksdale, “Inverted Pavement Response and Performance,” Transportation Research Record 1482, Transportation Research Board, National Research Council, Washington, D.C., 1995, pp. 102–110. Weingart, R., “Inverted Base: The Virginia Experience,” 2012 Transportation Research Board Mineral Aggregates Committee (AFP 70) Meeting Presentation, 91st Annual Meeting Transportation Research Board, Jan. 22–26, 2012, Washington, D.C. Wells, G.E., and R.C. Adams, “Increased Single-Lift Depths for Aggregate Base Course in Highway Construction: Sec- ondary Road 1117 Relocation, Richmond County, North Carolina,” presented at the 5th Annual Symposium of the International Center for Aggregate Research (ICAR), 1997. White, D.J., P. Vennapusa, and C.T. Jahren, Determination of the Optimum Base Characteristics for Pavements, Iowa DOT Project TR-482, Center for Transportation Research and Education, Iowa State University, Ames, 2004. Williamson, T.G., and E.J. Yoder, An Investigation of Com- paction Variability for Selected Highway Projects in Indi- ana, Publication FHWA/IN/JHRP-67/05, Joint Highway Research Project, Indiana Department of Transportation and Purdue University, West Lafayette, 1967. Womack, W.W., “Increasing the Lift Thickness for Aggregate Base: Virginia Department of Transportation Perspec- tive,” presented at the 5th Annual Symposium of the Inter- national Center for Aggregate Research (ICAR), 1997.