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.

101 chapter five COMPACTION, QUALITY CONTROL, AND FIELD PERFORMANCE INTRODUCTION This chapter presents detailed findings on different approaches used by transportation agencies for compaction testing on laboratory samples, field compaction, QC/QA, and field per- formance evaluations of constructed UAB/subbase layers. Different aspects of compaction and QC of UAB and sub- base construction are discussed by first introducing the theory of compaction along with the objectives behind compacting unbound aggregate pavement layers. This is followed by a review of different types of compactors commonly used for compacting UAB and subbase layers in the field. The concept of QC is introduced, emphasizing that constructed layer den- sity measurement is the most commonly used field evaluation tool for verifying the adequacy of UAB/subbase construc- tion. However, laboratory testing is needed to establish the target densities and acceptance criteria in field compaction of aggregate layers. Different field techniques used to measure densities of constructed pavement layers are discussed, with particular emphasis on the widespread nuclear gauge-based direct density measurement methods. The concept of modulus-based compaction control is intro- duced by highlighting its potential advantages, such as continu- ous compaction control for uniformity over the “spot checking” compared with density-based compaction control approaches. Different IC approaches are also discussed through a review of equipment manufacturers. Furthermore, experiences of differ- ent states in the United States for implementing IC approaches are presented, along with their preliminary findings. Finally, this chapter discusses other portable devices used for measuring the in situ moduli of constructed pavement layers. Salient features of each device are discussed, and the advantages and disadvantages of individual devices are high- lighted. Research studies and trial projects conducted by dif- ferent agencies through QC using these portable devices are listed and a summary of their important findings provided. COMPACTION AND QUALITY CONTROL Theory and Objectives of Compaction Compaction is defined as the densification of soils and con- struction materials through the application of mechanical energy. Primary objectives of compaction are to (1) reduce/ prevent detrimental settlements (compaction leads to better packing of individual particles, thus reducing the potential for excessive settlement); (2) increase the shear strength and thus improve slope stability; (3) improve the bearing capacity of pavement subgrades and granular subbase/base layers; and (4) control undesirable volume changes caused by frost action, swelling, and shrinkage (Holtz 1990). Although most initial research efforts focused on compaction were concerned with the compaction of soils (Proctor 1933; Seed 1959), the com- paction of aggregates as geomaterials is equally important in the construction of pavement layers. As discussed, the pri- mary mechanism of load transfer within an aggregate layer is through particle-to-particle interlock. The process of compac- tion reorients the particles within a loose aggregate layer and creates a densely packed matrix. This densely packed aggre- gate matrix demonstrates significantly higher shear strength and resilient modulus, as well as significantly lower suscep- tibility to permanent deformation compared with a loose uncompacted layer of aggregates. Although the process of compaction invariably results in higher densities achieved in the compacted layers, the achieve- ment of higher densities is not one of the primary objectives of compaction. Rather, density is an indicator of achieved compac- tion levels and often can be linked to other mechanical properties of soils and aggregates, such as shear strength and susceptibility to permanent deformation accumulation. Inadequate compac- tion of pavement subgrade, subbase, or base layers may result in excessive rutting, leading to pavement shear failure. Marek and Jones (1974) highlight the difference between “compaction” and “density,” emphasizing that two aggregate base materials compacted to the same density may be at com- pletely different stages of compaction. They emphasize that the state of compaction of an aggregate material is dependent on its gradation, so depending on the amount of fines in an aggregate matrix, higher density numbers may not always correspond to “better” states of compaction. According to Proctor (1933), the compactability of a soil or aggregate layer depends on the following factors: (1) compactive energy, (2) moisture content, and (3) soil/aggregate type. Establishing the Target Density for Field Compaction Control The primary method for measuring the compaction level in a pavement layer is by comparing the achieved field densities with reference target values determined for the same material in the laboratory. The in-place densities of constructed layers are subsequently expressed as percentages of these reference

102 densities established in the laboratory. Construction specifi- cations for pavement layers often require the achieved field densities to be higher than a certain specified percentage of this target density value. The applicability of relative com- paction values thus determined is dependent on the validity of the following two assumptions: (1) the material tested in the laboratory is identical to the field material in gradation and specific gravity, and (2) similar compactive energies are imparted to the material in the field, as well as in the labora- tory. Upon the violation of one or both of these assumptions, the calculated “percent compaction” becomes meaningless (Marek and Jones 1974). Some of the commonly used meth- ods for establishing the “target density” values of unbound aggregate materials in the field are discussed here. Compaction Using Drop Hammer Methods Drop hammer methods are the test methods most commonly used for establishing the compaction characteristics of soils and aggregates in the laboratory. Originally proposed by Proctor (1933), these methods involve the compaction of a representative portion of the material into a standard size mold using a rammer dropped from a fixed height. Depend- ing on the weight of the rammer and the drop height, the procedure is termed either a standard or modified compac- tion procedure. It is important to note that the rammer blows in Proctor’s method were specified as “firm strokes,” whereas the test methods currently used involve free fall of the drop hammer over a fixed height. Equipment specifications, test methodology, and material to be tested using these methods are described in standard specifications by ASTM and AASHTO. The standard method involves compaction of a representative portion of the aggregate material into a standard size mold (101.6-mm or 152.4-mm diameter) with a 24.5-N (5.5-lbf) rammer dropped from a height of 305 mm (12.0 in.). The modified compaction method involves a 44.48-N (10.0-lbf) hammer dropped from a height of 457.2 mm (18.0 in.). Spec- ifications for the standard compaction procedure have been provided as ASTM D 698 or AASHTO T 99, and those for the modified compaction procedure have been provided as ASTM D 1557, or AASHTO T 180. Note that the ASTM and AASHTO methods differ somewhat in the maximum size of aggregate particles that can be tested. Moreover, owing to the use of a heavier hammer and higher drop height, the modi- fied compaction procedure imparts much higher compaction energy to the aggregate specimen (4.5 times) than does the standard compaction procedure. Table 11 lists the similarities and differences between the two compaction methods as spec- ified by the ASTM and AASHTO standards. It is important to note that several state and Canadian provincial agencies use modified versions of the original ASTM and AASHTO spec- ifications as part of their agency guidelines. Although these agency-specific guidelines are somewhat different from the ASTM and AASHTO standards, the basic procedures and principles remain the same. Figure 78 shows the typical compaction curves for a com- monly used dense-graded crushed limestone material with 10% P200 fines. As shown in the figure, a higher compac- Equipment/Test Parameter ASTM AASHTO Standard (D 698) Modified (D 1557) Standard T 99 Modified T 180 Mold diameter Method A: 101.6 mm Method B: 101.6 mm Method C: 152.4 mm Method A: 101.6 mm Method B: 152.4 mm Method C: 101.6 mm Method D: 152.4 mm Mold volume (cm3) 943.0 for 101.6-mm diameter mold 2,124 for 152.4-mm diameter mold Number of layers 3 5 3 5 Number of blows/layer 25 for 101.6-mm diameter mold 56 for 152.4-mm diameter mold Material specifications [material finer than sieve opening size (%)] Method A: 4.75 mm Method B: 9.50 mm Method C: 19.0 mm Method A: 4.75 mm Method B: 4.75 mm Method C: 19.0 mm Method D: 19.0 mm TABLE 11 COMPArISON OF ASTM AND AASHTO TEST METHODS GOvErNING THE COMPACTION OF SOILS AND AGGrEGATES USING DrOP HAMMEr METHOD FIGURE 78 Typical compaction curves for a dense-graded crushed limestone material with 10% P200 fines (1 pcf = 16.02 kg/m3).

103 tive energy leads to an increase in the maximum dry density (MDD) value and a decrease in the OMC. Note that drop-hammer–based compaction methods are commonly used by transportation agencies to establish refer- ence target densities, such as 95% to 100% of laboratory MDD values, before the construction of UAB and subbase layers. The survey of state and Canadian provincial transportation agencies conducted under the scope of the current synthesis study indicated that 42 of 46 responding agencies use drop- hammer–based methods to establish the compaction char- acteristics of unbound aggregate materials in the laboratory. Only two agencies (the Kansas and Alabama DOTs) reported the use of vibratory compaction methods. One agency (Alberta Transportation, Canada) does not require aggregate compac- tion characteristics to be established in the laboratory. It is important to note that drop-hammer–based compaction methods specified in AASHTO T 99 and T 180 were derived from the original methods proposed by Proctor (1933), which in turn were developed for fine-grained soils. Accordingly, the use of impact compaction may not be adequate for establishing the compaction characteristics of certain aggregate types, such as open-graded materials with insufficient P200 fines. Absence of sufficient P200 fines results in “shifting” of individual aggre- gate particles under impact compaction, thus preventing the formation of a densely packed matrix. Thus, vibratory com- paction can be used to establish the compaction characteristics of such materials. Although ASTM method D 7382 (Standard Test Methods for Determination of Maximum Dry Unit Weight and Water Content Range for Effective Compaction of Granu- lar Soils Using a vibrating Hammer) provides such an alterna- tive, no AASHTO method directs the compaction of unbound aggregates using vibratory methods. Because the compaction of UAB/subbase layers commonly involves vibratory and shearing action, establishing aggregate compaction character- istics in the laboratory using vibratory or gyratory compaction methods may lead to better representation of field conditions in the laboratory. Control Strip or “Test Strip” Method The control strip or “test strip” technique involves the con- struction of a control strip using the same material as that used to construct the UAB/subbase layer. This strip is compacted through repeated rolling and vibration, and density tests are performed after each rolling until no additional increase in den- sity is noticed. The average final density of the control strip is used as the “maximum” density for the particular aggregate material. Construction specifications require the aggregate base/subbase layer to be compacted to a certain percentage of this “maximum” density (Marek and Jones 1974). For successful implementation of the control strip method, the compaction of the strip is correlated to previously established compac- tion results. In the absence of adequate moisture and compaction equipment, the maximum density achieved in the strip may not represent the densest feasible state of compaction. A new “test strip” often is required when (1) a change in the source of material is made; (2) a change in the material from the same source is observed; and/or (3) when 10 test sections are approved with- out a new control strip (Anday and Hughes 1967). Solid Volume Density Method Certain construction specifications can also be based on the solid volume density of the aggregate as a reference. The solid volume density is obtained by multiplying the specific gravity of the aggregate material with the unit weight of water (9.81 kN/m3 or 62.4 pounds per cubic foot). The solid volume density represents the density of a particular aggregate material in a “void-less” matrix. Constructed layer densities are expressed as a percentage of the solid volume density, and the fraction is termed relative solid density. One of the most common examples of construction specifications using the solid volume density method can be seen in the construction of G1 base in South Africa. Note that for successful implementation of the solid volume density method, the correlation between achieved densities in the field and the void-less density should be known. For example, a relative solid density value of 86% typically corresponds to approximately 100% to 105% of the maximum dry density value obtained using the modified compaction method as per AASHTO T 180 (Buchanan 2010). Note that this correlation is just an example, and the exact correlation will vary depending on the aggregate mineralogy, gradation, and particle shape and surface texture. Key Lessons • Compaction characteristics of aggregates estab- lished in the laboratory are strongly governed by compaction methods. For example, the maximum dry density values established using AASHTO T 99 are consistently lower than those established using AASHTO T 180 because of the lower compaction energy imparted to the aggregate specimen in the former. • Drop-hammer–based compaction methods (e.g., AASHTO T 99 and T 180) may not be adequate for coarse-grained aggregates, particularly with low fines (P200) contents. • Test procedures similar to ASTM D 7382 that establish the moisture-density curves for un bound aggregates using a vibratory (or a gyratory) com- pactor may lead to better representation of field conditions in the laboratory. Compaction Variables and Equipment Types The DOC achieved in a constructed unbound aggregate layer is dependent on the interaction between several vari- ables, which can be broadly classified into the following two

104 categories: (1) aggregate material and layer characteristics and (2) compaction equipment and operating characteris- tics. Different variables falling under the two categories are described here. Aggregate Material and Layer Characteristics The following variables can be grouped under this category and affect the DOC of unbound aggregate layers by governing the arrangement of individual particles in the aggregate matrix: 1. Type of parent rock (in terms of the hardness and dura- bility of individual particles); 2. Particle shape and surface texture; 3. Gradation or particle size distribution; 4. Construction lift thickness; 5. Moisture content; and 6. Layer support conditions. Compaction Equipment and Operating Characteristics The following variables related to the compaction equipment and operating characteristics affect the DOC achieved in unbound aggregate layers by governing the amount of energy imparted to the layer surface: 1. Roller type; 2. Roller weight/energy; 3. Roller speed or dwell time; 4. Number of passes or coverages; 5. Rolling zone; and 6. Rolling pattern. The roller types commonly used in the compaction of con- structed pavement layers are discussed here. Smooth Drum Rollers Smooth drum rollers are probably the most commonly used compaction devices during the con- struction of UAB and subbase layers. These rollers can con- sist of a single drum or dual drums that apply pressure across the drum width. These rollers can also be “static” or “vibra- tory” in nature. Static smooth drum rollers compact the pave- ment layers through static application of the equipment dead weight, but vibratory smooth drum rollers are equipped with oscillatory vibrators to increase the energy transmitted to the layer surface. vibratory smooth drum rollers are best suited for unbound aggregates and non cohesive soils. In addition, these rollers sometimes are used to finish subgrades before the construction of base/subbase layers. Figure 79 is a photo of a smooth drum vibratory roller (single drum) used to com- pact a crushed limestone base course. Sheepsfoot Rollers Also known as “studded rollers,” these typically are used in the compaction of cohesive soils. These rollers have a drum with several rounded or rectangular protru- sions or feet and apply very high contact pressures to the soil layer being compacted. The vertical contact stress is dependent on the spacing of the protrusions on the drum and creates a kneading action that compacts the layer “bottom up.” Once compaction is complete, the roller “walks out” of the lift, leav- ing the surface fairly rough. This kneading or shearing action maximizes a cohesive soil’s strength at high density levels. Some sheepsfoot rollers are equipped with oscillatory vibra- tors to increase the effectiveness across a broader range of soil (Christopher et al. 2010). One variation of the sheepsfoot roller, known as the tamping foot roller, has feet with sloping sides. Because of the sloping nature of the feet, tamping foot rollers leave the compacted layer surface fairly smooth. Figure 80 is a photo of a sheepsfoot roller used to compact a low plasticity clayey silt. Pneumatic or Rubber-Tire Rollers Pneumatic-tired roll- ers generally have two tandem axles with three to six wheels each. The wheels are arranged so that the rear ones will run in the spaces between the front ones, theoretically leaving no ruts. The weight of ballast carried by the equipment chassis can FIGURE 79 Compaction of a crushed limestone base course using a smooth drum vibratory roller. FIGURE 80 Compaction of a low plasticity clayey silt (CL-ML) using a sheepsfoot roller.

105 be varied to achieve the required compactive energy. Some- times the wheels are mounted slightly out of line with the axle, giving them a weaving action and the name “wobble wheel.” This condition improves the kneading action on the layer being compacted. These rollers often are used as an alternative for compacting a variety of soil types and are particularly effec- tive for noncohesive silty soils. Construction vehicles such as loaded dump trucks also can be used to serve as pneumatic roll- ers, especially during the placement of embankments. Pneu- matic rollers compact the soil layers top-down, and the zone of influence is relatively shallow, particularly for small-tire units (Ingersoll-Rand 1984). Impact Rollers Impact rollers comprise triangular ellipsoids or hexagonal drums to apply impact energy on to the layers being compacted as the roller moves along. Owing to the high impact energies being applied to the layer surface, these rollers achieve compaction at a faster rate and have a greater zone of influence compared with conventional smooth drum or sheeps- foot rollers. Although the use of impact rollers is common in Europe and South Africa, their availability in the United States is limited. Figure 81 is a photo of an impact roller. Grid Rollers Grid rollers have a cylindrical heavy steel surface consisting of a network of steel bars forming a grid with square holes and may be ballasted with concrete blocks. Grid rollers provide high-contact pressure but lit- tle kneading action and are suitable for compacting most coarse grained soils (RDSO, 2005). Table 12 is borrowed from Christopher et al. (2010) and lists the compactor types for different soil types (Original source: Rollings and Rollings 1996). FIGURE 81 Impact roller (http://www.fhwa.dot.gov/engineering/ geotech/pubs/05037/08.cfm). Soil Type First Choice Second Choice Comment Rock fill Vibratory Pneumatic — Plastic soils, CH-MH (A-7, A-5) Sheepsfoot or pad foot Pneumatic Thin lifts usually needed Low-plasticity soils, CL, ML (A-6, A-4) Sheepsfoot or pad foot Pneumatic, vibratory Moisture control often critical for silty soils Plastic sands and gravels, GC, SC (A-2-6, A-2-7) Vibratory, pneumatic Pad foot — Silty sands and gravels SM, GM (A-3, A-2-4, A-2-5) Vibratory Pneumatic, pad foot Moisture control often critical Clean sand, SW, SP (A-1-b) Vibratory Impact, pneumatic — Clean gravels, GW, GP (A-1-a) Vibratory Pneumatic, impact, grid Grid useful for over- sized particles Source: Rollings and Rollings (1996). TABLE 12 RECOMMENDED FIELD COMPACTION EQUIPMENT FOR DIFFERENT SOILS Key Lesson The use of roller types that are most suitable for the particular material types is critical to ensuring ade- quate compaction of unbound aggregate pavement layers. Measuring In-Place Density of Constructed Unbound Aggregate Layers Different methods exist for determining the moisture con- tent and achieved density of constructed UAB and subbase layers. Some of these methods are listed in Table 13 along with the test methods governing their respective procedures. Moisture Measurement Soil moisture measurements are routinely conducted during pavement construction for QA purposes. State construction guidelines typically specify methods, such as oven or hot plate drying or nuclear density gauge testing. Under some cir- cumstances, these methods may not be reliable, may require special licensing, and may be time consuming. Develop ing a performance-based specification that relates soil moisture to modulus requires that a rapid and reliable measure of unbound pavement material moisture content be developed. The key issues to consider are (1) type of moisture content measured (gravimetric or volumetric), (2) accuracy, (3) dura- bility, (4) response time, and (5) ease of use.

106 Direct Methods for Measuring Moisture Content The oven dry method, the microwave oven method, the direct heating method, and the calcium carbide gas pressure tester method (“speedy moisture content”) are examples of methods used to make gravimetric moisture measurements during pavement construction. Oven dry and direct heating methods operate on the principle that the water mass is the difference between the weights of the wet and oven dry samples. The soil water con- tent is expressed by weight as the ratio of the mass of water present to the dry weight of the soil sample. The field moisture oven has been used to measure moisture content during pave- ment construction (Camargo et al. 2006; White et al. 2009). The advantages of these devices are their ease of use and rela- tive inexpensive cost. Possible drawbacks are that their use can be time consuming, may require a large power source in the field, and requires an accompanying density test to convert to volumetric water content. Indirect Methods for Measuring Moisture Content Indirect methods for measuring volumetric water content rely on an empirically derived calibration with a measured variable such as dielectric permittivity. Dielectric methods have been used extensively for measuring soil water content in agricultural and geotechnical engineering applications. Time domain reflec- tometers, frequency domain reflectometers, and capacitance probes use the principles of the matrix dielectric permittivity to indirectly measure the volumetric moisture content in the soil. Several studies document the use of dielectric methods for measuring and monitoring pavement layer water content (Janoo et al. 1994; Rainwater and Yoder 1999; Roberson 2007). Spe- cific field devices that have been used to measure water con- tent during pavement construction include the Percometer, the TRIME-EZ probe, and the Trident Moisture Meter (veenstra et al. 2005; Camargo et al. 2006). Recently, the DM600 Road- bed Meter was developed specifically for measuring water con- tent of pavement materials; however, the device has not been extensively tested in the field. The advantages of the dielectric methods are fast equilibrium measurement times, relatively accurate measurements, easy automation, and the elimina- tion of the need for a density measurement. Drawbacks of the method are the possibility of inaccuracies resulting from high clay content and soil salinity, lack of durability, soil-specific calibration may be needed for some instruments, and instru- ments can be relatively expensive. Nuclear gauge-based moisture-density measurements have been in common use for transportation agencies for the last three decades. Commonly referred to as “nuclear density gauges,” these devices can measure the wet density and moisture content of compacted soil and aggregate layers. The wet density of a layer is measured by detecting the suppression of gamma waves from a source rod lowered into the ground (direct transmission mode). In a second mode of operation (backscatter mode), the source rod is at the same level as the detector (not lowered into the pavement layer), and gamma rays from the source are “scattered back” from the compacted layer to the gauge. Note that the use of the backscatter mode usually is not recommended for determining the density of granular base/subbase layers because granular layers usually are porous, and the presence of large voids can significantly reduce the amount of gamma rays that get reflected back to be captured by the detector. A nuclear density gauge monitors the moisture content of constructed pavement layers using a strong neutron source that emits neutrons into the surface. These neutrons are reflected upon colliding with the hydrogen atoms (similar in size to the neutrons) present in water. The amount of reflected neutrons detected by the gauge can be used to estimate the moisture con- tent of the pavement layers. Specifications for determining the moisture content of soil and aggregate layers using nuclear den- sity gauges are provided in AASHTO T 310 and ASTM D 3017. CURRENT STATE OF THE PRACTICE Based on the survey conducted, Figures 82 to 85 review trans- portation agency practices related to field compaction and con- struction QC of unbound aggregate layers. More than 75% of Parameter to Be Determined Name of Method ASTM AASHTO Moisture content Gravimetric D 2216 T 265 Microwave D 4643 N/A Calcium carbide gas pressure test D 4944 T 217 Density Sand cone D 1556 T 191 Sand Replacement D4914 N/A Balloon D 2167 T 205a Oil or water Drive cylinder D 2937 T 204 a Moisture and density Rapid D 5080 N/A Nuclear Moisture D 3017 T 310 Density D 2922 Time domain reflectometry D 6780 N/A a N/A: Not available a Withdrawn from latest standards. TABLE 13 DIFFERENT METHODS TO DETERMINE THE MOISTURE DENSITY OF COMPACTED AGGREGATE BASE AND SUBBASE LAYERS IN THE FIELD

107 4% 30% 76% 13% (2) (14) (35) (6) 0 10 20 30 40 0% 20% 40% 60% 80% 100% According to quarry reported moisture content Sampled during construction/compaction for laboratory testing Measured through field testing using other methods (nuclear methods, etc.) Other (Contractor responsibility, etc.) Number of Responses Percentage of Survey Respondents 46 survey respondents FIGURE 82 Different methods used by transportation agencies to control the moisture content of constructed/compacted UAB layers in the field. 30% 7% 89% 7% 0% 9% (14) (3) (41) (3) (0) (4) 0 10 20 30 40 0% 20% 40% 60% 80% 100% Gradation Proof-rolling Measurement of constructed layer density Field measurement of constructed layer modulus Continuous compaction control by means of Intelligent Compaction (IC) equipment Other Number of Responses Percentage of Survey Respondents 46 survey respondents FIGURE 83 Primary approaches used by transportation agencies for evaluating degree of compaction and construction quality control of UAB/subbase layers. 89% 16% 2% 11% (39) (7) (1) (5) 0 10 20 30 40 0% 20% 40% 60% 80% 100% Nuclear density methods (ASTM D 2922 / AASHTO T310 or T238) Sand cone method (ASTM D 1556 / AASHTO T191) Balloon method (ASTM D 2167) Other (please indicate) Number of Responses Percentage of Survey Repondents 44 survey respondents FIGURE 84 Methods commonly used by transportation agencies for mea- suring constructed aggregate base/subbase layer densities in the field.

108 52% 37% 41% 44% 7% 41% 17% (24) (17) (19) (20) (3) (19) (8) 0 10 20 30 40 0% 20% 40% 60% 80% 100% Yes (please select all possible reasons) Safety Concerns Nuclear certification too expensive Nuclear certification too inconvenient Non-nuclear methods provide better results No Other Number of Responses Percentage of Respondents 46 survey respondents FIGURE 85 Responses to the question “Is there interest to implement nonnuclear density measurement methods for construction quality control of unbound aggregate base/ subbase layers.” maximum achievable densities in the laboratory through com- monly used compaction tests. Although research has success- fully correlated higher densities to unbound aggregate layer stiffness or resilient modulus improvements (Rowshanzamir 1995; Tutumluer and Seyhan 1998), M-E pavement design methods do not consider aggregate layer density as an input into pavement thickness design. The resilient modulus, on the other hand, governs the nature of stress dissipation in an aggre- gate layer because of wheel load, and thus is an essential input for mechanistic analysis of the layered pavement structure. This alone has made the alternative of measuring in situ layer modulus attractive for pavement designers, although a chal- lenging task now deals with how to develop related construc- tion specifications for field modulus control. Growing interest in modulus-based compaction control procedures has led to the development of several different alternatives for nondestructive field modulus measurements of pavement layers. von Quintus et al. (2009) and Puppala (2008) present an extensive overview of different techniques and devices available for the measurement of in-place pave- ment layer moduli. The underlying techniques used for in-place modulus measurement of UAB and subbase layers are listed in Table 11 along with examples of devices based on the cor- responding principles. Note that the devices listed in Table 11 all function based on different principles, so the reported values may have different dimensions. Some devices are based on the principle of measuring stiffness, whereas some measure modulus. It is important to note that “stiffness” is not an independent soil parameter and is dependent on the area over which the load is applied. However, “modulus” is truly an independent soil parameter and is independent of the compaction equipment. Thus, for true representation of com- pacted layer properties, a device should report the modulus value and not just the stiffness value (Briaud and Seo 2003). the responding agencies control moisture content of constructed/ compacted UAB in the field (see Figure 82). Field moisture and density measurements using nuclear density gauges is a com- mon practice in 89% of the responding agencies (see Figures 83 and 84). The in-place densities thus determined are compared with laboratory-established compaction characteristics to check the DOC achieved in constructed aggregate layers. Only 28% of the responding agencies construct test strips to establish roller patterns and check for compaction density growth of aggregate layers. More than 50% of the responding agencies expressed interest in implementing non-nuclear density measurement methods for construction control of UAB/subbase layers owing to one reason or another (see Figure 86). However, several of them indicated a lack of confidence in the performance of non-nuclear moisture-density measurement alternatives. Key Lessons • Measurement of compacted unbound aggregate layer density using a nuclear gauge is a common practice among transportation agencies. • There is growing interest among agencies in gradu- ally moving toward density measurement systems that are not nuclear based owing to certification and convenience issues associated with nuclear gauge testing. IN-PLACE MODULUS MEASUREMENT OF CONSTRUCTED AGGREGATE LAYERS Quality control and quality assurance (QC/QA) of constructed unbound aggregate pavement layers traditionally has been based on target density values, expressed with respect to the

109 Of the previously listed devices, the FWD is the most com- monly used device by transportation agencies for indirectly measuring (or backcalculating from measured deflections) the in-service pavement layer moduli. The survey of state and Canadian transportation agencies conducted under the scope of the current synthesis study indicated that 27 of 46 responding agencies use FWD testing to assess the struc- tural condition of UAB and subbase layers in existing pave- ment structures. However, although FWD testing on in-service pavement structures is a fairly common practice among trans- portation agencies, the use of an FWD device directly on top of UAB/subbase is a relatively new practice. For example, the UK performance-based specifications recommend the use of FWD to check the adequacy of constructed unbound aggregate layers (Interim Advice Note 2009). The most com- monly adopted techniques for checking the quality of con- structed unbound aggregate layers using in-place modulus measurements involve portable devices such as the LWD, GeoGauge, and surface seismic, or continuous measurement devices such as instrumented compactors. LWDs are used as primary field devices in several countries, including Ger- many, Austria, and Sweden, to measure earthwork stiffness/ modulus. However, in the United States, only Indiana DOT uses LWD to measure the modulus of constructed unbound aggregate layers. Several studies have been conducted focusing on the cor- relation between field measured stiffness/modulus to den- sity, correlation between stiffness/modulus values reported by different devices, and repeatability of values reported by individual devices. (Puppala 2008; von Quintus et al. 2009). Chen et al. (1999) conducted field modulus/stiffness tests on different subgrade and base materials in more than six Texas districts and made the following observations: • Field-measured density of constructed pavement layers was not sensitive to change in modulus. • Both the soil stiffness gauge (Humboldt GeoGauge) and seismic techniques, such as Dirt-Seismic Pavement Analyzer and Olson-SASW, reported modulus values that were consistent with those reported by conven- tional FWD and showed promise for being used as QC devices, Nazzal (2003) conducted extensive field testing to evalu- ate the potential of several NDT devices, such as the soil stiffness gauge (Humboldt GeoGauge), DCP, and LWD to measure the stiffness/strength parameters of highway mate- rials and embankment soils during and after construction. A strong correlation was found between layer modulus values reported by LWD and GeoGauge-type devices and those measured from conventional FWD testing. Furthermore, higher coefficients of variation were reported to be associ- ated with LWD-measured modulus values than were those measured by the GeoGauge, indicating the GeoGauge is a more “consistent” device (Nazzal 2003). von Quintus et al. (2009) reported that deflection-based methods such as LWD and FWD had limited potential for QC purposes. Testing several constructed pavement sections using different devices, they reported that deflection-based methods were not able to consistently identify areas with con- struction anomalies. Moreover, modulus values were influ- enced by the underlying layers, resulting in lower or higher and more variable modulus values. More recently, Mishra et al. (2012) measured the field modulus values of full-scale unsurfaced pavement test sections using a Dynatest LWD (Model 3031) and a soil stiffness gauge (Humboldt GeoGauge). By measuring the layer moduli on top of the prepared subgrade as well as constructed unbound aggregate layers using both devices, they reported that both devices were capable of identify- ing anomalies in construction conditions. Higher modulus values were measured by the soil stiffness gauge compared with the LWD because of the relatively smaller magnitudes of strains imposed on the pavement layers by the soil stiff- ness gauge when compared with the LWD. Similar to the findings of von Quintus et al. (2009), Mishra et al. (2012) reported that the LWD-measured modulus values were affected by layer thicknesses. Mooney and Miller (2009) also reported a depth of influence for LWD between 0.9 and 1.1 times the plate diameter, making it susceptible to the influences of underlying layers, especially for testing on thin aggregate layers. The soil stiffness gauge was found to be more consis- tent in measuring field modulus values irrespective of constructed layer thicknesses. Although von Quintus et al. (2009) reported a strong correlation between layer modulus values measured by a soil stiffness gauge and the achieved dry density values, Mishra et al. (2012) did not observe any such correlation from their testing. Moreover, several research studies have focused on the “validity” of stiffness/modulus values reported by these devices with respect to the actual stress-strain states experienced by pavement layers under traffic loading. For example, Mooney and Miller (2009) measured the in situ stress and strain behav- ior during LWD testing and showed that the LWD test engages a nonlinear soil modulus. Key Lessons • Several research and implementation projects have reported different degrees of success with in-place modulus measurement devices. • Although these devices have been used success- fully to identify anomalies in construction condi- tions, extensive calibration for local materials is needed before they can be used as primary tools for QC.

110 MODULUS-BASED COMPACTION CONTROL Need for Modulus-Based Compaction Control Although the measurement of the dry unit weight and mois- ture content of constructed UAB/subbase layers is relatively straightforward and practical, it does not provide any direct indication about the layer modulus or shear strength. Moreover, it is important to note that the same density can be obtained for at least two different moisture contents on either side of the compaction (moisture-density) curve. Thus, it is not ideal to use the achieved dry density as the only criterion for compaction/ construction QC. A modulus-based compaction control method combines the aspects of construction QC with in-place mea- surements of layer moduli. Desired Characteristics of a Modulus-Based Compaction Control System For developing a modulus-based construction specifica- tion, a few key issues must be properly considered, namely: (1) measurement depth, (2) induced stress state and stress path in relation to strength, and (3) use of proper algorithms for layer modulus estimation. Ideally, a field technique would estimate the elastic modu- lus of the individual pavement layers separately to be consis- tent with how material is represented in the MEPDG. Note that field devices also may provide stiffness measurements of aggregate materials belonging to depths that are often incon- sistent with layer thickness. This limitation for devices that measure deeper than the layer thickness can be overcome, as has been demonstrated in recent research by Senseney and Mooney (2010), who successfully extracted unbound layer moduli using the LWD with center position and radial offset sensors (similar to FWD). The approach is simple and robust for stiff-over-soft conditions (e.g., base over subbase, subbase over subgrade). If such techniques are not followed in the field, LWD moduli often will be dependent on depth of influ- ence but not exactly on the layer thickness of the constructed aggregate base/subbase. The way to control compaction is to ensure that the dry density is within tolerance from a target value, that the modulus is within tolerance from a target value, and that the water content is within tolerance of a target value. It is pos- sible to achieve reasonable control of compaction by ensuring that two of these three properties are within tolerance of their target values. In that respect, it is possible to control compac- tion by ensuring that the soil modulus and the water content are within tolerance of their target values. Implementing the modulus-based compaction control is desirable, but it can- not be used readily in practice because of the lack of proper guidelines and because specifications have not been estab- lished. Future practice no doubt will bring a basic need for the engineer to check that his or her modulus design assumption is verified in the field. CONTINUOUS COMPACTION CONTROL AND INTELLIGENT COMPACTION Continuous compaction control (CCC) uses vibratory drum compactors, which combine the technologies of a global positioning system (GPS), compactor-integrated measure- ment system, and an onboard display of real-time compac- tion measurements (Chang et al. 2011). Integration of these components allows compaction data, also known as roller integrated compaction measurements (RICMs), to be tied to a specific project location, which is constantly updated as the compaction progresses. CCC typically involves the use of vertical drum acceleration processed in the time and/ or frequency domains to assess the state of soil compac- tion. Early research in Sweden revealed that the vibration characteristics of the drum changed as the underlying soil was compacted (Thurner and Sandstrom 1980). For vibratory roller configurations, CCC involves measurement and anal- ysis of output from an accelerometer mounted to the roller drum and can provide a spatial record of compaction quality when linked to position measurements and a documentation system (Chang et al. 2011). Roller measurement values cal- culated based on accelerometer measurements use one of two different approaches: • Calculate a ratio of selected frequency harmonics for a set time interval, or • Calculate ground stiffness or elastic modulus based on a drum-ground interaction model and some assumptions. Continuous compaction control machines typically include the following (Peterson 2005): • Sensors to measure vibration of the drum; • Onboard electronics to record and process sensor output and record the stiffness; • Linkages to the machine controls to adjust compaction effort according to the measured stiffness; • Systems to record machine location; and • Either local storage or wireless communications sys- tems for data transfer. IC differs from CCC by providing real-time, automatic adjustment of compactor settings based on RICM values to ensure maximum compactor efficiency as compaction pro- gresses and soil properties change. The equipment adjust- ments based on RICM data generally involve modifying the eccentric mass moment with the drum(s) to affect excita- tion amplitude and frequency (Rinehart and Mooney 2008). Essentially, IC adds an additional feature over CCC by immediately interpreting RICMs and adjusting the compac- tor operating characteristics. A formal definition of IC has been given as: . . . the compaction of road materials, such as soils, aggregate bases, or asphalt pavement materials, using modern vibratory rollers equipped with an in situ measurement system and feed- back control. (http://www.intelligentcompaction.com/)

111 NCHRP Project 21-09, Intelligent Soil Compaction Sys- tems, listed the following as desirable features of an IC sys- tem: (1) continuous assessment of mechanistic soil properties (e.g., stiffness, modulus) through roller vibration monitoring; (2) automatic feedback control of vibration amplitude and frequency; and (3) an integrated global positioning system to provide a complete geographic information system-based record of the earthwork site. Five different types of RICMs [also known as intelligent compaction measurement values (ICMvs)] are used by com- monly available compaction equipment and typically vary from one equipment manufacturer to another. These five mea- surements can be broadly divided into three different theo- ries. Compaction meter value (CMv) and compaction control value are derived from amplitudes of the operating frequency and various harmonics and subharmonics. Roller-integrated stiffness (ks or kb) and vibration modulus (Evib) are based on measuring the soil displacement under a compactor-generated load. Finally, machine drive power, the relative newcomer to the group, is based on measuring the amount of power needed to propel the compactor over the soil. Chang et al. (2011) pres- ents a summary of the ICMvs in common use in the United States; these ICMvs are based on vibration frequency analysis or mechanical modeling (see Table 14). Need for Intelligent Compaction Intelligent compaction using CCC provides continuous data indicating the level of compaction achieved with every pass. Real-time processing of the data enables the equipment oper- ating characteristics to be changed frequently, thus imparting variable compactive energy levels to different spots as needed. This spontaneous adjustment of compactive effort has been found to be particularly important during the compaction of thick-lift aggregate layers. Evaluation of thick aggregate base materials in the United States has produced evidence to confirm the usefulness of this feature. This reduces the spatial variability associated with the DOC achieved in a given pavement layer. Continuous data collection and pro- cessing also eliminates the need for frequent “spot testing” for quality assurance process. Applying the optimum number of passes of the roller, an IC system significantly reduces the chances of over-compaction. Through comparison of data obtained from consecutive roller passes, an IC system can quickly identify “difficult to compact” areas, thus enabling field engineers to make decisions to remedy the problem. Moreover, continuous monitoring and compaction control can significantly reduce differential settlements that result from nonuniform compaction conditions in projects that rely solely on spot tests for QC. Finally, the level of compaction information gathered from rollers during the IC process is a better indicator of achieved compaction levels. This is primarily because of the significantly larger influence zones under a compactor compared with those corresponding to spot-testing equipment such as FWD, LWD, soil stiffness gauge, nuclear density gauge, or DCP. Chang et al. (2011) compares the influence zone under a roller to those under commonly used spot-testing devices (see Figure 86). In addition to the previously mentioned advantages, the following disadvantages of IC systems have been reported by researchers (Briaud and Seo 2003): 1. Requirement for sophisticated equipment in a rugged environment; 2. Requirement for operator training; and 3. More expensive than conventional compaction (may require an overall cost-benefit study). Test Category Underlying Principle Corresponding Devices Surface deformation Static load Benkelman beam Briaud compaction device (based on measuring the bending strain on a loading plate in contact with the ground) Steady state vibratory Soil stiffness gauge (e.g., Humboldt GeoGauge) Impact load Falling weight deflectometer (FWD) Portable falling weight deflectometer or light weight deflectometer (LWD) Sinusoidal load Dynaflect Road rater Continuous load Rolling wheel deflectometer Geophysical Wave propagation Ultrasonic body waves Ultrasonic surface waves Spectral analysis of surface waves (SASW) Multichannel analysis of surface waves Free-free resonant column tests Seismic pavement analyzer Portable seismic pavement analyzer TABLE 14 DIFFERENT METHODS AvAILABLE FOR IN-SITU MODULUS MEASUREMENT OF CONSTRUCTED PAvEMENT LAYERS

112 Synthesis of Past Research and Agency Experience with IC Systems Several research and trial projects have been conducted in the United States evaluating the application of IC systems as QC tools for UAB/subbase layer construction. Most notably, a Transportation Pooled Fund project, TPF-5(128) was conducted from 2008 to 2011 involving 12 participating state transporta- tion agencies (Georgia, Indiana, Kansas, Maryland, Minnesota, Mississippi, New York, North Dakota, Pennsylvania, Texas, Virginia, and Wisconsin). The primary objective of the study was to develop an IC expertise base, evaluate current IC equip- ment, and accelerate specification development. The follow- ing section presents a summary of some of the most notable findings from IC implementation studies in the United States. Minnesota Experience In 2005, Minnesota DOT used the MnROAD test track to demonstrate the Bomag system and other subgrade soil and aggregate base/subbase layer testing devices, including DCP, GeoGauge, and Light Weight Deflectometer (LWD), to deter- mine the relationship between the IC roller response output and independently measured soil properties (Peterson 2005). In general, the demonstration project concluded that CCC was an effective QC mechanism for soil compaction. Camargo et al. (2006) reported on a Minnesota case history involving IC equipment manufactured by Ammann, Bomag, and Caterpillar. Through compaction using IC equipment and spot-testing using QA devices such as DCP, LWD, and GeoGauge, it was observed that there was no significant dif- ference between the modulus measurements obtained from QA devices such as LWD or GeoGauge and the Bomag IC. Camargo et al. also highlighted the challenges associated with handling the massive amounts of data generated by IC equipment before IC specifications can be implemented for use by transportation agencies. Texas Experience Under the pooled fund study TPF-5(128), a field IC demon- stration was performed in Fort Worth, Texas, in 2008. The IC equipment used was Case/Ammann single-drum padfoot and smooth drum vibratory rollers and the Dynapac single-drum smooth drum vibratory roller. Using IC technology to compact cohesive subgrade and granular base layers, it was observed that in situ measurements using the calibrated moisture-density nuclear gauge, DCP, and LWD did not match well with those of the ICMVs. However, plate loading tests (PLTs) and FWD tests produced better correlation with the ICMVs (Chang et al. 2011) (see Table 15). NCHRP Project 10-65 NCHRP Project 10-65 (Von Quintus et al. 2009) used several different IC rollers (Bomag, Caterpillar, Case/Ammann) and FIGURE 86 Illustration of differences in measurement influence depths for different measurements (Chang et al. 2011).

113 an instrumented vibratory roller for the QC/QA of HMA mix- tures and unbound pavement layers. Through comparison of the IC response output parameters with modulus and density values measured using traditional as well as nondestructive testing devices, it was observed that the IC equipment was successful in detecting areas with significant density differ- ences. Moreover, the IC equipment output was found to cor- relate with other nondestructive testing density and modulus values (von Quintus et al. 2009). In summary, nearly all previous studies have concluded that the use of IC rollers has many advantages for use as a contractor’s QC tool to monitor the compaction of pave- ment materials and identify soft spots or weak areas along a project. Most studies have focused on the effect of increasing material compaction or density on the IC measured response and have reported good correlations between the IC output and density modulus for a specific material and project. Fewer studies have focused on the effect of temperature, moisture, material condition, and varying subsurface condi- tions on the responses and output from the IC measurement systems in terms of reducing the risk of making an incorrect decision during construction. Temperature of HMA, mois- ture content of unbound layers, and support conditions of the underlying layers are important factors related to the IC roller’s output. NCHRP Project 21-09 NCHRP Project 21-09, “Intelligent Soil Compaction Systems,” evaluated the reliability of different IC measurement systems and developed construction specifications for the compaction of subgrades, embankments, and UAB/subbase layers. Upon investigation of four vibration-based roller measurement val- ues (Mvs); (Ammann and Case/Ammann ks, Bomag Evib, Dynapac CMVD, and Sakai continuous compaction value), the study confirmed the dependence of roller Mv on the amplitude and frequency of roller vibration, and thus recommended con- struction specifications that allow IC during compaction but do not permit its use during roller-based QA. The construction specifications developed through this project were grouped into the following three categories (Mooney et al. 2010): 1. Option 1: This option uses CCC to identify weak spots in a compacted area to be further tested using commonly used spot tests. 2. Option 2: This option is based on statistical change in the roller Mv during compaction. It can be based on monitoring the difference between mean roller Mv from one pass to the other or on the percentage change in spatial roller Mv. Note that neither of these options requires the calibration of roller Mv using test strips. 3. Option 3: This option was further subdivided into alternatives that required the calibration of roller Mv with spot testing results and thus involves significant initial investments. Detailed discussion of these alter- native specifications and the challenges associated with the implementation of each can be found elsewhere (Mooney et al. 2010). Wisconsin Experience von Quintus et al. (2010) collected information and data on the use of IC technology to help Wisconsin DOT assess the validity and accuracy of IC in pavement construction. Through data collection from demonstration projects, they identified the following two usage areas as more mature and ready to have immediate positive benefits, especially for unbound materials: (1) use of IC rollers as a testing device to identify areas with weak supporting areas through con- tinuous mapping of the stiffness, and (2) development of stiffness-growth relationships to determine the rolling pat- tern and number of passes to achieve a specific stiffness level. They also recommended additional pilot projects to IC Measurement Units IC System Model Definition Compaction meter value (CMV) None Caterpillar Dynapac 2ACMV C A Machine drive power None Caterpillar Compaction control values None Sakai ' g AMDP P Wv Sin mv b g 0.5 1.5 2 2.5 3 0.5 100A A A A ACCV A A Stiffness (Kb) MN/m Ammann/ Case 2 0 0 cos b d d m ek m z Vibration modulus (Evib) MN/m 2 Bomag 31 2 2 2 2 2 1 2.14 0.5 ln 1 16 2 vib vib b e E aF z a E dm m m g Source: Chang et al. (2011). TABLE 15 SUMMARY OF IC MEASUREMENTS

114 increase contractor and agency personnel’s confidence in using the IC technology. Details on several other IC implementation projects were provided in Chang et al. (2011). Overall, all the IC imple- mentation studies have shown promising results regarding the potential of this technology to be used for QC purposes. Quality Assurance Specifications Based on Continuous Compaction Control White et al. (2007) conducted three field studies to inves- tigate the correlation of CMV (also known as Caterpillar compaction value) and machine drive power values from Caterpillar rollers and kB stiffness from Ammann rollers with in situ test measurements such as dry unit weight, DCP index, Clegg Impact Value, and LWD modulus and made the following observations: • The Ammann kB value showed a strong correlation with in situ test results for strips with a relatively wide range of material stiffnesses and a relatively weak correlation for strips with more uniform conditions. White et al. were also able to correlate the Ammann kB with rut depth measured after test rolling procedures. • IC technology could be successfully applied by the MnDOT as the principal QC tool on a grading project near Akeley, Minnesota. The entire project passed the test rolling acceptance criteria. Figure 87 shows the relationships between average in situ properties and RICM values as reported by White and Thompson (2008). Rinehart et al. (2009) compared the in situ stress states and stress paths experienced by a soil beneath two IC roll- ers on instrumented vertically homogeneous embankment soil and on layered base over subgrade to the stress states applied during AASHTO T 307 resilient modulus testing. Measuring the stress states to a depth of 1 m below the roller wheel, they observed that stress fields varied sig- nificantly with depth for the homogeneous embankment and FIGURE 87 Relationships between average in situ and RICM values (White and Thompson 2008).

115 the layered base over subgrade. The measured deviator stress (q) and mean stress (p) were observed to decrease by factors of 4 and 6, respectively, within the 300-mm thick crushed base. Even for low excitation forces applied by the rollers, the estimated q values in the crushed base were as much as 2.5 times greater than the maximum q values applied during resilient modulus testing in the laboratory. Mean stress (p) values observed in the field were approxi- mately 30% to 50% of the values applied during MR testing in the laboratory. Rinehart et al. (2009) also highlighted that the roller- based stiffness values determined during IC of layered constructions, such as pavement base over subgrade, are complex functions of material properties, layer thicknesses, and stress-dependent modulus parameters. Thus, any modulus- based compaction control protocol would be able to extract individual layer properties from these complex stiffness val- ues. This task may become simpler in the case of homoge- neous embankments with homogeneous modulus fields with depth. Current Specifications Based on Roller Integrated Compaction Measurements Specifications for IC were first introduced in Austria in 1990 (later modified in 1993 and 1999), Germany in 1994 (updated in 1997), Sweden in 1994 (revised in 2005), and Switzerland in 2006. The International Society of Soil Mechanics and Geotechnical Engineering (ISSMGE) adopted the Austrian specifications in 2005. In the United States, a pilot specifica- tion was developed by MnDOT in 2007 and updated in 2010. Similarly, a special specification was developed by Texas DOT in 2008, and a list of approved IC rollers was released in 2009. In July 2011, FHWA released generic specifications for compaction of soils and subbases using IC techniques. These generic specifications are to be modified by individ- ual agencies to meet specific requirements (www.intelligent compaction.com). The current IC specifications follow one of the following two approaches (www.intelligentcompaction.com): • Use of RICM values to identify weak spots, which can then be assessed by spot-testing techniques, such as moisture-density tests, PLTs, and/or LWD tests. Acceptance of the constructed layer is dependent on these “weak” spots satisfying minimum thresh- olds with respect to PLT modulus, LWD modulus, or density. This is the only approach permitted in Sweden. • Use of test beds to develop correlations between Mvs to PLT modulus, LWD modulus, or density in a defined calibration area. If a suitable correlation is found to exist, a target roller Mv is determined from the correla- tion and used for QA purposes. The Austrian/ISSMGE and German specifications each permit either of the two previously mentioned alternatives for construction QA using continuous compaction control. Based on a survey of European practice, Mooney et al. (2010) reported that implementation of the calibration approach was challenging and required a high level of on-site knowledge. Ongoing Effort: NCHRP Project 10-84 With an objective to develop modulus-based compaction con- trol specification in the United States, a research study funded by NCHRP is being undertaken. The developed specification shall (http://apps.trb.org/cmsfeed/TRBNetProjectDisplay. asp?ProjectID=2908): 1. Be based on field measures of the stiffness or modulus and moisture content of the compacted earthwork and unbound aggregate that can be correlated with design modulus values; 2. Provide a single, straightforward, and well-defined method for determining stiffness or modulus that is compatible with a variety of earthwork and unbound aggregate design methodologies; 3. Directly account for the seasonal variation of the mod- ulus of the compacted earthwork or unbound aggre- gate as the means to determine specification criteria and limits for compaction; 4. Use available models, devices, and methods, as defined in the current literature, including NCHRP Synthesis 382: Estimating Stiffness of Subgrade and Unbound Materials for Pavement Design (Puppala 2008); and 5. Be founded on a comprehensive review of the current literature on the long-term behavior of various soils and unbound aggregates in terms of the principles of unsaturated soil mechanics. The research work under the scope of this project will be conducted in three phases that have been subdivided into a total of 11 tasks. The project is currently in Phase III, with interim reports for Phases I and II already submitted to NCHRP. Current State of the Practice From the survey of state and Canadian provincial agencies conducted under the scope of this synthesis, it was observed that only one agency (Texas) has actively implemented IC techniques to construct in-service pavements with UAB/ subbase layers and has developed a specification for this purpose. Two agencies (Indiana and Georgia) implement modulus-based compaction control for the construction of UAB/subbase layers but only in demonstration projects. Indiana DOT reported the use of LWDs for field modulus measurement in demonstration projects.

116 CONSIDERATION OF SUCTION EFFECTS IN LAYER MODULUS ESTIMATION Background The modulus (and, correspondingly, the load response linked to performance) of earthwork and unbound aggregates is strongly influenced by the seasonal variation of their mois- ture content. This variation depends on material composition, DOC, and available free moisture, which is controlled primar- ily by the local climatic environment and the distance from the GWT. In developing a modulus-based construction speci- fication for compaction of earthwork and unbound aggregate that will provide a direct link with design parameters, all three factors—material, compaction, and moisture—should be examined on the basis of the principles of unsaturated soil mechanics with respect to highway engineering and construction. Unbound aggregate pavement layers are usually compacted at moisture contents corresponding to 80% to 90% saturation conditions (Gupta et al. 2007) and thus fall in the unsaturated regime. The unsaturated conditions and distribution of pore structure within the compacted aggregate layers lead to the development of negative pore water pressure (matric suction), which increases the effective stress states within the layers. This increase in the overall stress states may have a signifi- cant influence on the shear strength and stiffness (or modulus) of stress-dependent unbound aggregate materials. Moreover, suction conditions significantly affect soil volume change, the coefficient of permeability, and freeze-thaw susceptibility. Thus, a suitable procedure for evaluating constructed aggregate layer conditions includes the effects of matric suction. Within an unsaturated soil mechanics framework, soil suction can be represented as an independent stress state variable, and resilient modulus (MR) can be represented as shown in Equation 18 (Gupta et al. 2007): M k p k p p kR a b a k a k = −    +       1 6 7 3 2 3σ τoct   + α ψ β 1 1 18m ( ) where sb is the summation of all around or bulk stress; toct is the octahedral shear stress, ym is the matric suction, a1 and b1 are empirical fit parameters relating the resilient modulus (MR) to ym; pa is the atmospheric (normalizing) pressure; and k1, k2, k3, k6, and k7 are model parameters obtained from regression analyses. A modulus-based QA specification includes field moisture target values to ensure permanent deformation in the field remains below the allowable limit. A modulus specification based on unsaturated soil mechanics uses the volumetric moisture content (q) instead of the gravimetric value (w) (Gupta et al. 2007). A modulus-based compaction control specification incorporating soil suction effects also incor- porates methods to measure the soil suction of constructed pavement layers. In addition, the detrimental effects of soil wetting and the resulting loss of suction with changes in aggregate layer capillary structure should be considered in a comprehensive modulus-based QA specification. Methods for Measuring Soil Suction Different methods available to measure soil suction in the field can be broadly divided into two categories: (1) direct methods and (2) indirect methods (Lu and Likos 2004; Munoz-Carpena 2009). Oven dry methods and the calcium carbide gas pressure tester method (“speedy moisture content”) provide a direct measure of the gravimetric moisture content (w), whereas indirect methods for measuring q, such as time domain reflec- tometry, rely on an empirically derived calibration with a mea- sured variable, such as dielectric permittivity. Although many of the available devices have been used in pavement research, an evaluation of such devices for routine field use and spe- cifically for use in the development of a performance-based construction specification is still needed. Indirect methods for soil suction measurement include thermal conductivity methods and the filter paper method. Thermal conductivity methods have been used in pavement engineering research to characterize soil suction in the base and subgrade layers (Nichol et al. 2003; Roberson 2007). The measurement of soil suction is based on the theory that thermal conductivity properties of a soil are indicative of the soil water content. Soil suction is inferred by measuring the dissipation of heat within the sensor, which is related to the water content of the sensor that is in equilibrium with soil water (Roberson and Reece 1993; Reece 1996). The advantages of the thermal conductivity method are that it has Key Lessons • Continuous compaction control using different roller measurement values can significantly reduce spatial variability in compaction levels and can reduce the potential for differential settlements in constructed pavement layers. • Most research and implementation projects conducted in the United States involving the use of continuous compaction control and IC to construct UAB/subbase layers have reported considerable success. How- ever, such practices are not common for transpor- tation agencies. Encouraging more implementation projects across agencies can help to incorporate continuous compaction control and IC into agency practice. • Target relative stiffness values established during con- tinuous compaction control vary significantly from one compactor to another. For example, the compaction measurement value established using a double-drum IC roller is significantly different than that established using a single-drum IC roller.

117 a wide measurement range, is easily automated, and is not affected by salinity. The limitations include hysteresis, indi- vidual sensor calibration, and long equilibrium times. Filter Paper Methods Both the contact and noncontact filter paper methods are used to indirectly measure soil suction by measuring the amount of moisture transferred from a soil sample to a calibrated piece of filter paper. The filter paper is placed in direct contact with a soil specimen or is suspended (noncontact) over the soil specimen. Once equilibrium between the soil sample and the filter paper is reached, the water content of the filter paper is determined gravimetrically and related to the soil suction by means of a calibration curve particular to the type of filter paper (Lu and Likos 2004). The filter paper method is simple and relatively inexpensive. Drawbacks of the method are long equilibration times (6 to 10 days). of Intelligent Compaction Technology for Embankment Subgrade Soils, Aggregate Base, and Asphalt Pavement Materials, Federal Highway Administration, Washington, D.C., 2011. Chen, D.H., W.W. Wu, R. He, J. Bilyeu, and M. Arrelano, “Evaluation of In situ Resilient Modulus Testing Tech- niques,” Geotechnical Special Publication, No. 89, 1999, pp. 1–11. Christopher, B.R., C. Schwartz, and R. Boudreau, Geotech- nical Aspects of Pavements, Report FHWA NHI-10-092, National Highway Institute, Federal Highway Adminis- tration, Washington, D.C., 2010. Gupta, S.C., A. Ranaivoson, T. Edil, C. Benson, and A. Sawangsuriya, Pavement Design Using Unsaturated Soil Technology, MN/RC-2007-1, Minnesota Department of Transportation, St. Paul, 2007. Gupta, S.C., A. Singh, and A. Ranaivoson, Moisture Retention Characteristics of Base and Subbase Materials, Report submitted to the Minnesota Department of Transportation, St. Paul, Dec., 2004, p. 47. Holtz, R.D., Guide of Earthwork Construction, State of the Art Report 8, Transportation Research Board, National Research Council, Washington, D.C., 1990. Ingersoll-Rand, Compaction Data Handbook, Ingersoll-Rand Construction and Mining, 1984. Interim Advice Note 73/06 Revision 1 (2009): Design Guid- ance for Road Pavement Foundations (Draft HD25), Department of Transport, London, U.K., 2009. Lu, N. and W. Likos, Unsaturated Soil Mechanics, Wiley and Sons Inc., Hoboken, N.J., 2004, pp. 294–295. Marek, C.R. and T.R. Jones, Jr., “Compaction-An Essential Ingredient for Good Base Performance,” Proceedings from Conference on Utilization of Graded Aggregate Base in Flexible Pavements, Sponsored by National Crushed Stone Association, National Sand and Gravel Association, and National Slag Association, Mar. 25–26, 1974, Oak Brook, Ill. Mishra, D., E. Tutumluer, M. Moaveni, and Y. Xiao, “Labora- tory and Field Measured Moduli of Unsurfaced Pavements on Weak Subgrade,” Proceedings of the ASCE Geo-Institute GeoCongess, State of the Art and Practice in Geotechnical Engineering, Geotechnical Special Publication No. 225, CD-ROM, R.D. Hryciw, A. Athanasopoulos-Zekkos, and N. Yesiller, Eds., Oakland, Calif., Mar. 25–29, 2012. Mooney, M.A. and D. Adam, “vibratory Roller Integrated Measurement of Earthwork Compaction: An Overview,” Proceedings, FMGM2007—International Symposium on Field Measurements in Geomechanics, Sep. 24–27, 2007, Boston, Mass. Mooney, M.A. and P.K. Miller, “Analysis of Lightweight Deflectometer Test Based on In situ Stress and Strain Response,” Journal of Geotechnical & Geoenvironmental Engineering, ASCE, vol. 135, No. 2, 2009, pp. 199–208. Mooney, M.A., R.v. Rinehart, N.W. Facas, O.M. Musimbi, D.J. White, and P.K.R. vennapusa, NCHRP Report 676: Intelligent Soil Compaction Systems, Transportation Research Board of the National Academies, Washington, D.C., 2010. Key Lesson Suction effects and resulting changes in aggregate layer modulus should be considered during the design of UAB/subbase layers. REFERENCES Anday, M.C. and C.S. Hughes, “Compaction Control of Gran- ular Base Course Materials by use of Nuclear Devices and a Control Strip Technique,” Highway Research Record 177, Transportation Research Board, National Research Council, Washington, D.C., 1967, pp. 136–143. Briaud, J.L. and J. Seo, Intelligent Compaction Overview and Research Needs, Report to the Federal Highway Adminis- tration, Washington, D.C., Texas A&M University, College Station, Tex., 2003. Briaud, J.L. and K.Y. Rhee, The BCD: A New Instrumenta- tion for Compaction Control, Final Report for IDEA Proj- ect 118, Transportation Research Board of the National Academies, Washington, D.C., 2009. Buchanan, S., Inverted Pavements-What, Why, and How? AFTRE Industry Education Webinar, Aggregates Founda- tion for Technology, Research, and Education, Alexandria, va., June 1, 2010. Camargo, F., B. Larsen, B. Chadbourn, R. Roberson, and J. Siekmeier, “Intelligent Compaction: A Minnesota Case History,” 54th Annual University of Minnesota Geo- technical Conference, Feb. 17, 2006 [Online]. Available: http://www.intelligentcompaction.com/downloads/Papers Reports/MnDOT_Camargo_IC%20A%20Minnesota%20 Case%20History_2006.pdf. Chang, G., Q. Xu, J. Rutledge, B. Horan, L. Michael, D. White, and P. vennapusa, Accelerated Implementation

118 Munoz-Carpena, R., “Field Devices for Monitoring Soil Water Content,” University of Florida IFAS Extension, 2009. Nazzal, M., Field Evaluation of In situ Test Technology for QC/QA Procedures during Construction of Pavement Layers and Embankments, MS Thesis, Louisiana State University, Baton Rouge, 2003. Nichol, C., L. Smith, and R. Beckie, “Long-term Mea- surement of Matric Suction Using Thermal Conductiv- ity Sensors,” Canadian Geotechnical Journal, vol. 40, 2003, pp. 587–597. Peterson, D.L., Continuous Compaction Control: MnRoad Demonstration, Report No. MN/RC-2005-07, Office of Materials and Road Research, Minnesota Department of Transportation, Maplewood, 2005. Proctor, R.R., “Fundamental Principles of Soil Compaction,” Engineering News Record, vol. 7, Aug.–Sep. 1933. Puppala, A.J., NCHRP Synthesis 382: Estimating Stiffness of Subgrade and Unbound Materials for Pavement Design, Transportation Research Board of the National Academies, Washington, D.C., 2008, 137 pp. Research Designs and Standards Organization (RDSO), Study Report on Compaction Equipment and Construction Machinery, Report No. GE-R-76, Geotechnical Engineer- ing Directorate, RDSO, Manak Nagar, Lucknow, India, 2005. Reece, C.F., “Evaluation of a Line Heat Dissipation Sensor for Measuring Soil Matric Potential,” Soil Science Society of America Journal, vol. 60, 1996, pp. 1022–1028. Richter, C.A., Seasonal variations in the Moduli of Unbound Pavement Layers, Publication No. FHWA-HRT-04-079, Federal Highway Administration, Washington, D.C., 2006. Rinehart, R.v. and M.A. Mooney, “Instrumentation of a Roller Compactor to Monitor vibration Behavior during Earthwork Compaction,” Automation in Construction, No. 17, 2008, pp. 144–150. Rinehart, R.v. and M.A. Mooney, “Measurement of Roller Compactor Induced Triaxial Soil Stresses and Strains,” Geotechnical Testing Journal, ASTM, vol. 32, No. 4, 2009. Rinehart, R.v., M.A. Mooney, and J.R. Berger, “Comparison of Stress States and Paths: vibratory Roller-Measured Soil Stiffness and Resilient Modulus Testing,” Trans- portation Research Record: Journal of the Transporta- tion Research Board, No. 2116, Transportation Research Board of the National Academies, Washington, D.C., 2009, pp. 8–15. Rinehart, R.v. and M.A. Mooney, “Measurement Depth of vibratory Roller-Measured Soil Stiffness,” Géotechnique, vol. 59, No. 7, 2009, pp. 609–619. Roberson, R.L. and C.F. Reece, “Comparison of Heat Dissi- pation Sensor, Tensiometer, and Thermocouple Psychrom- eter Response to Changing Soil Moisture Conditions,” Poster/Abstract 85th Annual Meeting Soil Science Society of America, 1993. Roberson, R., Impact of Pavement Drainage Design on Material Stiffness, M.S. Thesis, University of Minnesota, 2007. Rollings, M.P., and R.S. Rollings, Geotechnical Materials in Construction, McGraw–Hill, N.Y., 1996. Rowshanzamir, M.A., Resilient Cross Anisotropic Behav- ior of Granular Base Materials under Repetitive Load- ing, Ph.D. Dissertation, University of New South Wales, Australia, 1995. Seed, H.B., “A Modern Approach to Soil Compaction,” Pro- ceedings of the Eleventh California Street and Highway Conference, Institute of Transportation and Traffic Engi- neering, University of California, 1959, pp. 77–93. Senseney, C.T. and M.A. Mooney, “Characterization of Two- Layer Soil System Using a Lightweight Deflectometer with Radial Sensors,” Transportation Research Record: Journal of the Transportation Research Board, No. 2186, Transportation Research Board of the National Academies, Washington, D.C., 2010, pp. 21–28. Thurner, H.F. and A. Sandstrom, “Continuous Compaction Control, CCC,” Proceedings of the International Confer- ence on Compaction, Paris, France, 1980, pp. 237–245. Tutumluer, E. and U. Seyhan, “Neural Network Modeling of Anisotropic Aggregate Behavior from Repeated Load Triaxial Tests,” Transportation Research Record 1615, Transportation Research Board, National Research Coun- cil, Washington, D.C., 1998, pp. 86–93. von Quintus, H.L., C. Rao, R.E. Minchin, S. Nazarian, K.R. Maser, and B. Prowell, NCHRP Report 626: NDT Technol- ogy for Quality Assurance of HMA Pavement Construction, Transportation Research Board of the National Academies, Washington, D.C., 2009. von Quintus, H.L., C. Rao, H. Titi, B. Bhattacharya, and R. English, Evaluation of Intelligent Compaction Technol- ogy for Densification of Roadway Subgrades and Struc- tural Layers, Final Report: WHRP Study #00092-08-07, Wisconsin Department of Transportation, Madison, 2010. White, D.J., M. Thompson, and P. vennapusa, Field Valida- tion of Intelligent Compaction Monitoring Technology for Unbound Materials, Report No. MN/RC-2007-10, Center for Transportation Research and Education, Iowa State University, Ames, 2007. White, D.J., and M.J. Thompson, “Relationships between In Situ and Roller-Integrated Compaction Measurements for Granular Soils,” Journal of Geotechnical and Geoenviron- mental Engineering, vol. 134, No. 12, 2008, pp. 1763–1770. White, D.J., P. vennapusa, J. Zhang, H. Gieselman, and M. Morris, Implementation of Intelligent Compaction Performance Based Specifications in Minnesota, Report No. MN/RC 2009-14, Minnesota Department of Transpor- tation, St. Paul, 2009.