Mechanistic-Empirical Pavement Design Procedure For .

List of FiguresFigure 2.1 Kelvin and Maxwell Model representation for viscoelastic materials. 20Figure 2.2 a Typical subgrade and base course interface before migration of finesor penetration of aggregate (after Valentine, 1997). . 34Figure 2.2 b Typical subgrade and base course interface showing surface ruttingas a result of the migration of fines and penetration ofaggregate (after Valentine, 1997). . 34Figure 2.3 Idealized surface load stress distribution through drained aggregatebase course and saturated aggregate base course (after Cedergren,1977). 36Figure 2.4 Percent of design aggregate thickness lost as a function ofsubgrade CBR (after FHWA, 1989). . 37Figure 2.5 Resilient modulus verses bulk stress for various fine contents (afterJorenby and Hicks, 1986). 52Figure 2.6 Resilient modulus verses percent added fines (after Jorenby andHicks, 1986). . 53Figure 2.7 Design criteria for unreinforced pavement section thickness versusequivalent reinforced thickness (after Webster, 1991). . 58Figure 2.8 Pavements cross sections. . 61Figure 3.1 Original road alignment and new alignment. . 72Figure 3.2 Layout of the test sections and support structures. . 73Figure 3.3 Plans for service pole. 83Figure 3.4 Plans for top half of bunker - exploded view. 85Figure 3.5 Cross-section of bunker placement. . 86Figure 4.1 A schematic of the Kulite type 0234 earth pressure cell. . 93Figure 4.2 A schematic of the Carlson type TP-101 earth pressure cell. . 93Figure 4.3 A schematic of the Kyowa type embedded HMA strain gage. . 96Figure 4.4 A typical cross-section of PVC conduit from junction box to road. . 100Figure 5.1 FWD-derived surface moduli. 118Figure 5.2 Apparent subgrade resilient modulus variation over time. . 122Figure 5.3 Flowchart of the iterative procedure of transition layer thicknessdetermination. . 123xi

Figure 5.4 Development of transition layer. 124Figure 5.5 Location of GPR passes. . 127Figure 5.6 GPR scans over section 1 (June, 1997). . 128Figure 5.7 GPR scans over section 2 (June, 1997). . 129Figure 5.8 GPR scans over section 3 (June, 1997). . 129Figure 5.9 Pressure at the top of the subgrade for 80 kN axle load and 560 kPatire pressure with HMA temperature of 25 C. . 131Figure 5.10 ELSYM 5 master curve for 80 kN axle load, 560 kPa tire pressure at12 C. 135Figure 5.11 Contour plot for vertical stress. 138Figure 5.12 Contour plot for vertical deformation. . 139Figure 5.13 Contour plot for horizontal strain. . 140Figure 5.14 (a) Excavated section with location of forms. . 143Figure 5.14 (b) Cross-sectional view of the excavated section. 143Figure 6.1 Laminate Geometry and In-plane forces . 154Figure 6.2 Shift factors for the HMA samples collected from the test section. . 160Figure 6.3 Moving Load variation as a function of time for 550 kPa appliedpressure and 150 mm tire radius. . 163Figure 6.4 Physical representation of conversion of static response to dynamicresponse. . 164Figure 6.5 A schematic pavement cross-section with critical locations. . 165Figure 6.6 Extent of vertical plastic strain accumulation for HMA samples overtime. . 167Figure 6.7 Percentage plastic strain over time (35 kPa). . 168Figure 6.8 Percentage plastic strain over time (105 kPa). . 168Figure 6.9 Pavement life as a function of material constant “m”. . 171Figure 6.10 Proposed mechanistic pavement design approach . 172Figure 6.11 Secondary and interstate pavement systems with mechanical loading. . 173Figure 6.12 Maximum, minimum and base temperature over the span of 12seasons (months). 174Figure 6.13 Number of trucks per day passing over the interstate and thesecondary road. . 175Figure 6.14 Average speed of the vehicles over the span of 12 seasons for thetest problem. . 175xii

Figure 6.15 Rut depth for the secondary road. . 177Figure 6.16 Plastic strain at the top of the HMA layer. 177Figure 6.17 Plastic strain at the bottom of HMA layer. 178Figure 6.18 Rut depth for the Interstate road. . 180Figure 6.19 Plastic strain at the top of the HMA layer. 181Figure 6.20 Plastic strain at the bottom of HMA layer. 181xiii

1 INTRODUCTION1.1 BackgroundDuring the past two decades, the use of geosynthetics in pavements has increaseddramatically (Barksdale et al., 1989; Dass, 1991; Austin and Coleman et al., 1993;Koerner et al., 1994; Al-Qadi et al., 1994, 1996, 1997). Various studies have beenperformed in the past few years to validate the performance of geosynthetics in highwaypavements (Li et al., 1992; Al-Qadi et al., 1994; Koerner and Koerner, 1994). Attemptswere made to develop design methods for pavements stabilized with geotextiles andgeogrids (Hass, 1986; Carroll et al., 1987; Barksdale et al., 1989; Koerner et al., 1994),but with little success. Several field and laboratory studies using static and dynamicloading conditions were conducted to validate the various claims of improving pavementperformance due to geosynthetic inclusion.In 1993, researchers at Virginia Techundertook a laboratory study to validate the performance of geogrids and geotextilesunder controlled laboratory conditions using dynamic loading (Al-Qadi et al., 1994; Smithet al., 1995). The conclusions from that study supported the idea that geotextiles doimprove flexible pavement performance due to the separation mechanism they introducein a layered system, and not by reinforcement, as previously believed (Koerner andKoerner 1994; Al-Qadi et al., 1994; Koerner, 1994; Carroll et al., 1987; Smith et al.,1995; Jorenby et al., 1986; Lair and Brau, 1986).To validate this finding, acomprehensive field study was conducted.In June 1994, a 150-m-long secondary road flexible pavement section in BedfordCounty, Virginia was instrumented. The test section, which consists of aggregate baseand hot-mix asphalt (HMA) layers, is part of the realignment of intersection of Routes757 and 616. The pavement section is located on a curve of constant radius with anintersection at mid-length.The average daily traffic on the pavement section isapproximately 530 (700 in summer) vehicles, with approximately 8% trucks.Thepavement section is composed of nine individual sections, each 15-m-long. Sections onethrough three have a 100 mm-thick limestone base course (VDOT 21-B); sections fourthrough six have a 150 mm-thick base course; sections seven through nine have a 200mm-thick base course. Three sections were stabilized with geotextiles and three withgeogrid. The other three were kept as control sections. Geosynthetic stabilization wasplaced at the base course-subgrade interface. One of each stabilization category was1

included in each base course thickness group. The non-stabilized sections served asthe basis of comparison of the benefits of incorporating geosynthetics in the flexiblepavement system.The outside wheel path of the inner curve lane is instrumented with strain gages,pressure cells, surface piezoelectric sensors, thermocouples, and moisture sensors. AKeithley 500-A data acquisition system was used to collect instrument responses onsite. Data was collected and analyzed from the various instruments over three years. Inaddition to instrument responses from the pavement, laboratory testing on specimensobtained from the field was conducted, such as resilient modulus, creep compliance,dynamic accumulated plastic strain for HMA, and resilient modulus and characterizationof base course and subgrade. Rut measurements, ground penetrating radar survey, andFalling Weight Deflectometer (FWD) tests were also performed on the test section. Thecollected data were analyzed to determine sectional performance and the effectivenessof geosynthetics at the base course-subgrade interface.1.2 Objective and ScopeThe objective of this study was to validate the Virginia Tech laboratory investigationas to the geosynthetic effectiveness in flexible pavements and to determine the extent ofits benefit. It is an important objective of this study to quantify this effectiveness andprovide a better understanding of the geosynthetic mechanism in pavements. In order toachieve these objectives, the test section in Bedford County, Virginia was selected,because, it presented a freshly constructed test section with optimal characteristics tostudy short-term and long-term performance of geosynthetically stabilized pavements inthe field. Approximately 150 instruments of various functions were used to monitor thevehicular and environmental effects on the pavement.In addition to subjecting the test section to accelerated heavy vehicular loading,periodic performance evaluations were conducted, including measuring subgraderesilient modulus using FWD and rut measurements. The collected data were analyzedand correlated to sectional performance.This step involved data reduction, whichincluded sorting and analysis of raw data. A set of computer programs to visually reducethe raw data were developed for this purpose.The FWD data was also used toquantitatively estimate the development of a “transition layer” at the base coursesubgrade interface.2

The measured field data were also used to calibrate and check a pavement designmethod developed under this project. This method is transient elastic viscoelasto-plasticmethod and uses mechanical and thermal loading to determine cumulative strains anddeformations in the HMA and subgrade.1.3 Report StructureThis report addresses field testing of geosynthetically stabilized pavement andincludes site selection, construction, instrumentation, field testing, and data analysis. Italso discusses the development and testing of a mechanistic pavement design approachthat considers geosynthetics in the design.Chapter 2 discusses state-of-the-art knowledge in the area of pavements, whichcovers pavement analysis and design, geosynthetic stabilization and case studies ofvarious projects involving the use of geosynthetics and instrumentationChapter 3 presents details of the site and the pavement construction process. It alsoincludes details on the material characterization and pre-construction field testing.Chapter 4 discusses the instrumentation types, calibration, installation, and sectionalperformance.Chapter 5 details the data collection from instruments under normal and controlledtraffic, rut depth, FWD measurement, and ground penetration radar (GPR) survey. Italso presents a comparison of subgrade and base course material gradations after threeyears of in-field testing and monitoring.Chapter 6 discusses new pavement analysis and design approach including basicequations and overall formulation. This method allows the inclusion of geosynthetics inthe design procedure.Chapter 7 presents the findings and conclusions of this study.3

2 PRESENT STATE OF KNOWLEDGEThis section outlines the development of pavement design and research from theearliest times to the present. It proceeds from the nominal through the empirical and thesemi-empirical, and finally to the semi-mechanistic approaches.2.1 Pre-Modern Pavement DesignThe earliest examples of highways and pavements have been identified in ancientBabylon and Egypt. However, little is known of these structures, because the ravages oftime and environment have left few for modern examination. The greatest of the ancientroad-builders were without doubt the Romans, who had established an efficient highwaysystem throughout their empire by the year 200 BC. This system was established topermit the rapid movement of military forces and support materials anywhere within theimperial boundaries. Later, the same system was used to facilitate the administration ofgovernment throughout the empire and to enhance the system of commerce needed tosupport such a large economy.The Roman concept of pavement massive at very best, required a vast amount oflabor and materials. The construction was based upon erecting a pavement structureover the existing ground of such strength that all conceivable loads could be carried withminimum post-construction maintenance. Typically, a foundation of large stones wasplaced over the natural ground, and a matrix of finer stones and limestone dust wasplaced over them.Finally, a surfacing of carefully dressed hard stone provided awearing surface. These structures were frequently as thick as one meter, and thereforewell-raised above the surrounding topography, which mitigated against poor drainage.The height of these structures above the natural ground indeed provides the origin of theterm “highway.”There is little to term “design” in this process of early road building: these pavementswere constructed “by the book,” an archetypal military manual.However, it isrecognized that early field engineers did have to exercise judgment in selecting theappropriate structure in consideration of two situations: available materials and groundsupport. A number of different cross-sections (Viae terrenae, Viae glaraetae, and Viaemunitae) are reported in various archaeological excavations and in contemporaryliterature (Caesar, 1996; Vitruvius, 1960). Of particular note are the pavement structuresused in Roman Britain to carry highways over marshy or boggy areas. Mats of logs and4

planking were first laid down (sometimes supported by timber piles) and bound intoflexible rafts, over which the more traditional solid Roman pavements were constructed - the first “raft foundations.”That many of these pavement structures were over-designed for the traffic andenvironments under which they were to operate is evidenced by the fact that many arestill in existence, and indeed many form part of existing highway systems (e.g., the M-1Motorway in UK follows much of the alignment of “Watling Street,” a Roman road, andfrequently incorporates the original Roman pavement within the modern infrastructure).Following the fall of the Roman Empire (circa 400 AD), the highway system fell intodisrepair, and as Europe fractionated into small, independent fiefdoms, parts wereactually destroyed to prevent the easy movement of armies an

2.2 Modern Pavement Design 5 2.2.1 Early Modern Pavement Design ( 1775 to 1900 AD) 5 2.2.2 Modern Pavement Design (20th Century) 7 2.3 Flexible Pavement 8 2.3.1 Methods Based on Soil Properties 8 2.3.2 Performance-Based Pavement Design Methods 10 2.3.3 Empirical-Mechanistic Methods 15 2.3.4 Other Atte