@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Applied Science, Faculty of"@en, "Civil Engineering, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Sigurdsson, Oddur"@en ; dcterms:issued "2008-08-21T21:48:01Z"@en, "1993"@en ; vivo:relatedDegree "Master of Applied Science - MASc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description "A full scale field trial was carried out to investigate the performance of different geosynthetics in unpaved road construction over soft ground. The test site comprises five 16 m long, by 4.5 m wide test sections, built on a subgrade of undrained shear strength approximately 40 kPa. One is unreinforced and serves as a control section in the study, three sections include a geotextile, and one includes a geogrid. Each test section incorporated a variable thickness of sandy gravel base course material, between 25 and 50 cm thick. They were trafficked in sequence by a vehicle of standard axle load. An important governing parameter for interpretation of behavior is the influence of base course thickness on the relationship between number of passes and rut depth. Performance of the test sections was evaluated from measurements of rut depth, base course thickness, base course deformations, geosynthetic strain, and deformed profile of the geosynthetic, with increasing number of vehicle passes. Vehicle trafficking was continued to a rut depth of about 20 cm, which constitutes a serviceability failure. Results from the full scale field trial show a better performance in the reinforced sections than the unreinforced section. The performance of the unreinforced section shows good agreement with other well-documented field data at large rut depths, between 10 and 15 cm, but not at small ruts. Although the four geosynthetics exhibited a broad range of stiffness and material properties, the general performance of the four reinforced sections was similar on the thicker base course layers. This is attributed to a reinforced mechanism governed by stiffness and separation, and all materials appear adequately stiff for the site condition and vehicle loading. On the thinner subgrades, a tensioned-membrane effect is mobilized, and a significant difference is observed between the geosynthetics."@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/1462?expand=metadata"@en ; dcterms:extent "6131031 bytes"@en ; dc:format "application/pdf"@en ; skos:note "Geosynthetic Stabilization of Unpaved Roads on Soft Ground :a Field EvaluationbyOddur SigurdssonB.Sc., The Technical College of Iceland, 1991A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIES(DEPARTMENT OF CIVIL ENGINEERING)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAOctober 1993© Oddur Sigurdsson, 1993In presenting this thesis in partial fulfilment of the requirements for an advanced degreeat the University of British Columbia, I agree that the Library shall make it freelyavailable for reference and study. I further agree that permission for extensive copyingof this thesis for scholarly purposes may be granted by the head of my department orby his or her representatives. It is understood that copying or publication of this thesisfor financial gain shall not be allowed without my written permission.(signatureDepartment of 6-1\\ii Evzi■ The University of British ColumbiaVancouver, CanadaDate 01'ci- (5 —(-k I i AbstractA full scale field trial was carried out to investigate the performance of differentgeosynthetics in unpaved road construction over soft ground. The test site comprisesfive 16 m long, by 4.5 m wide test sections, built on a subgrade of undrained shearstrength approximately 40 kPa. One is unreinforced and serves as a control section inthe study, three sections include a geotextile, and one includes a geogrid. Each testsection incorporated a variable thickness of sandy gravel base course material, between25 and 50 cm thick. They were trafficked in sequence by a vehicle of standard axleload. An important governing parameter for interpretation of behavior is the influenceof base course thickness on the relationship between number of passes and rut depth.Performance of the test sections was evaluated from measurements of rut depth, basecourse thickness, base course deformations, geosynthetic strain, and deformed profileof the geosynthetic, with increasing number of vehicle passes. Vehicle trafficking wascontinued to a rut depth of about 20 cm, which constitutes a serviceability failure.Results from the full scale field trial show a better performance in the reinforcedsections than the unreinforced section. The performance of the unreinforced sectionshows good agreement with other well-documented field data at large rut depths,between 10 and 15 cm, but not at small ruts. Although the four geosyntheticsexhibited a broad range of stiffness and material properties, the general performance ofthe four reinforced sections was similar on the thicker base course layers. This isattributed to a reinforced mechanism governed by stiffness and separation, and allmaterials appear adequately stiff for the site condition and vehicle loading. On thethinner subgrades, a tensioned-membrane effect is mobilized, and a significantdifference is observed between the geosynthetics.iiTable of ContentsAbstract^ iiTable of Contents^ iiiList of Tables viList of Figures^ viiAcknowledgment xiCHAPTER 1 Introduction^ 11.1 INTRODUCTION TO GEOSYNTHETICS^ 11.1.1 Geotextiles^ 21.1.2 Geogrids 31.1.3 Geomembranes 31.1.4 Geonets^ 41.1.5 Geocomposites 51.2 FUNCTIONS AND APPLICATIONS OF GEOSYNTHETICS^ 51.2.1 Separation^ 51.2.2 Reinforcement 61.2.3 Filtration 81.2.4 Drainage^ 91.2.4 Low Permeability Barrier^ 101.3 SCOPE OF RESEARCH^ 121.4 RESEARCH OBJECTIVES 131.5 OUTLINE OF THESIS^ 13CHAPTER 2 Introduction to Geosynthetics^ 152.1 INTRODUCTION^ 152.2 FUNDAMENTALS OF DESIGN APPROACHES^ 16iiiTable of Contents (cont.)2.2.1 Bearing Capacity^ 162.2.2 Membrane Effect 222.2.3 Effect of Anchorage 252.3 UNPAVED ROAD CONSTRUCTION^ 272.3.1 Field Studies^ 272.3.2 Model Studies 302.3.3 Analytical Studies 302.4 REVIEW OF DESIGN PROCEDURES^ 322.4.1 Empirical Design Approaches 322.4.2 Semi-Theoretical Design Methods 35CHAPTER 3 Test Site Description^ 543.1 GEOLOGICAL HISTORY 543.1.1 Geological History of the Lower Mainland^ 543.1.2 Surficial Geology of Lulu Island^ 563.2 DESCRIPTION AND LOCATION OF TEST SITE^ 573.3 SUBGRADE SOIL PROPERTIES^ 593.3.1 Site Stratigraphy^ 593.3.2 Soil Properties 613.3.3 Shear Strength Characteristics^ 65CHAPTER 4 Description of Test Sections 694.1 TEST SECTION LAYOUT^ 694.1.1 Arrangement of the Test Sections^ 704.1.2 Instrumentation and Measurements 714.1.3 Construction of the Test Section 744.2 BASE COURSE MATERIAL PROPERTIES^ 764.3 GEOSYNTHETIC PROPERTIES^ 774.4 LOADING VEHICLE^ 79ivTable of Contents (cont.)CHAPTER 5 Field Data^ 805.1 INTRODUCTION 805.2 FIELD DATA^ 825.2.1 Surface Profiles^ 825.2.2 Rut Development 865.2.3 Subgrade Profiles and Settlements^ 955.2.4 Geosynthetic Strain Measurements 101CHAPTER 6 Interpretation of Field Data^ 1076.1 INTRODUCTION^ 1076.2 INFLUENCE OF THE GEOSYN'THETICS^ 1076.3 UNREINFORCED PERFORMANCE 1196.4 REINFORCED PERFORMANCE^ 1216.5 LOAD DISTRIBUTION ANGLE 1236.6 STRESSES AND BEARING CAPACITY FACTORS^ 124CHAPTER 7 Conclusions and Recommendations^ 1077.1 CONCLUSIONS^ 1297.2 IMPLICATIONS FOR ENGINEERING DESIGN^ 132Bibliography^ 135vList of TablesTable 3.1 Specific Gravity Results^ 61Table 4.1 Properties of the geotextiles 77Table 4.2 Properties of the geogrid^ 78Table 5.1 Summary of geosynthetic strain measurements^ 105viList of FiguresFigure 2.1Figure 2.2Figure 2.3Figure 2.4Figure 2.5Figure 2.6Figure 2.7Figure 2.8Figure 2.9Figure 2.10Figure 2.11Figure 2.12Failure of soil beneath a continuous smooth and rough footing ^ 17Membrane action of geosynthetic in unpaved road^ 22Membrane effect in unpaved roads with geosynthetics - basicparameters^ 23Assumed deformed shape of base course-subgrade interface^ 37Load distribution by base course layer:(a) Case without geotextile^ 39(b) Case with geotextile 39Assumed parabolic shape of deformed geotextile^ 41Road cross section showing pertinent factors 46Location of the plastic zones^ 48Bearing capacity of subgrade 51Load spread angle below plain strain footing^ 51Soil block in equilibrium analysis^ 52Normal and shear stress interaction at subgrade-basecourse interface^ 52Figure 3.1Figure 3.2Figure 3.3Figure 3.4Figure 3.5Figure 3.6Figure 3.7Figure 3.8Index map showing location of Fraser Lowland^ 55Location of the research site^ 58General research site details 58Location of tests^ 60Typical test pit results 60Particle size distribution - Hydrometer analysis^ 62Atterberg limits and indices - Test results 63Relationship between liquid limit and plasticity index for thesubgrade soil^ 64viiList of Figures (cont.)Figure 3.9 Field vane shear test results^ 67Figure 4.1Figure 4.2Figure 4.3Figure 4.4Figure 4.5Figure 4.6Test section layout 71Test section geometry^ 71Instrumentation details 72Settlement plate - Details^ 74Base course aggregate grain size distribution^ 76Configuration of the loading vehicle 79Figure 5.1Figure 5.2aCross section numbers in each test section^ 81Surface profiles at cross section 5, h = 25 cm,test sections 1 to 5 ^ 83Surface profiles at cross section 3, h = 35 cm,test sections 1 to 5 84Figure 5.2c Surface profiles at cross section 1, h = 50 cm,test sections 1 to 5^ 85Figure 5.3a Rut depth versus number of passes - Section 1 raw data^ 88Figure 5.3b Average rut depth versus number of passes -Unreinforced data^ 88Figure 5.4a Rut depth versus number of passes -Section 2 raw data 89Figure 5.4b Average rut depth versus number of passes -Texel Geo 9 data^ 89Figure 5.5a Rut depth versus number of passes - Section 3 raw data^ 90Figure 5.5b Average rut depth versus number of passes -Polyfelt TS 700 data^ 90Figure 5.6a Rut depth versus number of passes - Section 4 raw data^ 91Figure 5.2bviiiList of Figures (cont.)Figure 5.6b Average rut depth versus number of passes -Tensar BX 1100 data^ 91Figure 5.7a Rut depth versus number of passes - Section 5 raw data^ 92Figure 5.7b Average rut depth versus number of passes -Polyfelt TS 600 data^ 92Figure 5.8 Subgrade profiles in test sections 2.5 and 2.3 -Texel Geo 9 96Figure 5.9 Subgrade profiles in test sections 3.5 and 3.3 -Polyfelt TS 700^ 96Figure 5.10 Subgrade profiles in test sections 4.6 and 4.3 -Tensar BX 1100 97Figure 5.11 Subgrade profiles in test sections 5.5 and 5.3 -Polyfelt TS 600^Figure 5.12 Subgrade settlement - Section 1 ^Figure 5.13 Subgrade settlement - Section 2^Figure 5.14 Subgrade settlement - Section 3^Figure 5.15 Subgrade settlement - Section 4^Figure 5.16 Subgrade settlement - Section 5^Figure 5.17 Geotextile strain measurements - Section 2Figure 5.18 Geotextile strain measurements - Section 3Figure 5.19 Geogrid strain measurements - Section 4 Figure 5.20 Geotextile strain measurements - Section 5979999100100101102103103104Figure 6.1 Comparison of average rut depth versus number of passes:h = 25 cm^ 110Figure 6.2 Comparison of average rut depth versus number of passes:h = 30 cm 110ixList of Figures (cont.)Figure 6.3Figure 6.4Figure 6.5Figure 6.6Figure 6.7Figure 6.8Figure 6.9Figure 6.10Figure 6.11Figure 6.12Figure 6.13Figure 6.14Figure 6.15Figure 6.16Figure 6.17Figure 6.18Comparison of average rut depth versus number of passes:h = 35 cm^ 111Comparison of average rut depth versus number of passes:h = 40 cm 111Comparison of average rut depth versus number of passes:h = 50 cm^ 112Subgrade settlements versus number of passes:h = 25, 35 & 50 cm^ 113Changes in base course thickness with number of passes^ 114Rut depth versus base course thickness - Section 1^ 116Rut depth versus base course thickness - Section 2 117Rut depth versus base course thickness - Section 3^ 117Rut depth versus base course thickness - Section 4 118Rut depth versus base course thickness - Section 5^ 118Unreinforced data comparison^ 120Reinforced data comparison 122Load distribution angle versus base course thickness^ 123Vertical stresses on subgrade surface versus base coursethickness^ 125Bearing capacity factor prediction - Unreinforced sections^ 127Bearing capacity factor prediction - Reinforced sections^ 127Figure 7.1 Base course thickness versus number of passes - Costsavings estimation^ 134AcknowledgmentsI wish to thank my supervisor, Dr. R.J. Fannin, for his invaluable support,encouragement, patience, and discussion throughout the research. My thanks are alsoextended to Dr. P.M. Byrne for his additional review of the thesis.The project was supported by the Natural Sciences and Engineering ResearchCouncil of Canada (NSERC), through the Collaborative Research and Development(CRD) Grant program. The industrial partners were Polyfelt Inc., Texel Inc., andTensar Earth Technologies Inc. Additional support from the Public RoadAdministration of Iceland is acknowledged. The test site and the loading vehicle wereprovided by the Geotechnical and Material Engineering section of B.C. Ministry ofTransportation and Highways, and their in-kind contribution is gratefullyacknowledged. I would also like to thank my friends and fellow students in thegeotechnical research group and the Department of Civil Engineering for theirfriendship and cooperation during my graduate studies.On a more personal note, I would like to thank my wife Nanna, my son Siggi,and my parents Hrefna and SigurOur, who encouraged me to undertake this endeavourand supported me throughout.xiChapter 1Introduction1.1 INTRODUCTION TO GEOSYNTHETICSGeosynthetics is a relatively new term in the engineering practice, used for arange of synthetic materials which have become increasingly important in a widevariety of engineering applications. As the name indicates, geosynthetics are artificialmaterials, the vast majority of which are made from polymers. Polymeric materials arecomposed of many small parts, called monomers, in which homopolymers are the mostcommon type of monomer used in the manufacture of geosynthetics. Polymers arealso made of the so called copolymer and terpolymer which describe two and threerepeating units in a chain, respectively. The synthetic materials made from polymersare non-biodegradable and are collectively called geosynthetics. They are furtherbroken into the following categories: geotextiles; geogrids; geomembranes; geonets;and geocomposites; which are described in some detail in the next section.1Chapter 1 - Introduction^ 21.1.1 GeotextilesGeotextiles are the oldest type of geosynthetics. The main polymeric materialsused in the manufacture of geotextiles are polypropylene and polyester which arethermoplastic materials, though polyamide and polyethylene are also used(Koerner, 1990). The manufacture of geotextiles begins with the production of fiberfilaments by extruding melted polymers through a spinneret. The monofilament fibersare then obtained by hardening the fiber filaments and stretching them in order toreduce the diameter and improve the strength of the fiber. There are other types offibers used in manufacturing of geotextiles, the multifilament which is produced bycombining the monofilament, and staple fibers which are obtained by cutting filamentsinto a short lengths of 2 to 10 cm. These staples can then be twisted into staple yarns.Slit films which are typically 1 to 3 mm wide are another fiber type used in theconstruction of geotextiles. The slit films are produced by slitting an extruded plasticfilm with blades. Slit film yarn is then made by combining the slit film fibers.In the geotextile manufacturing process, these fibers are converted into a planarpermeable structure called a fabric. The types of fabrics most often used forgeotextiles are wovens and nonwovens. Knitted fabrics are also available, but arerarely used in the geotextile industry.The woven geotextiles are made of two perpendicular sets of parallel filaments orstrands of yarn systematically interlaced using regular textile weaving machinery toform the planar structure. There are a wide variety of fabric weaves for use asgeotextiles, but they are kept relatively simple for the conventional industrial fabrics.The manufacture of nonwoven geotextiles is different from that of wovens, andmore complicated. Nonwoven geotextiles are formed by arranging the fibers in anoriented or random pattern to form a continuous loose web. The fabric mat is thenChapter 1 - Introduction^ 3bonded together by using one or a combination of the following processes: thermal ormelt-bonding, where heat causes partial melting of filaments or fibers at their crossoverpoints; chemical or resin-bonding, where an acrylic resin is either sprayed on the looseweb or used to impregnate the filaments and fibers; mechanical bonding orneedle-punching process, where barbed needles are punched through the loose web inorder to entangle the fibers.1.1.2 GeogridsGeogrids are more recent than geotextiles in the field of engineering and aremade of high-modulus polymer materials. The polymeric materials used in themanufacturing of geogrids are high-density polyethylene and high coherence andtoughness polyester or polypropylene. Extruded sheets are either drawn or rolled,under controlled strain rates and temperature, in one or two perpendicular directionsdepending on which style is preferred. In the drawing or rolling process, the smallregular pattern perforations become large quasi-rectangular openings. Geogrids areavailable in both uniaxial and biaxial styles which is dependent on how the heavy gaugepolymer sheets are drawn after being perforated.1.1.3 GeomembranesGeomembranes form the second largest group of geosynthetic materials and aremostly made of semicrystalline thermoplastic resins such as polyethylene, polyvinylchloride and chlorosulfonated polyethylene. The manufacturing of the three basicgeomembrane types begins with production of the raw materials, which includes thepolymer resin together with various additives used to obtain the final very lowpermeability thin plastic or rubber sheet product.Chapter 1 - Introduction^ 4A single-layer, unreinforced geomembrane is the simplest production type, wherethe raw materials are compounded before the extrusion process. In the extrusionprocess the extruder melts down the raw materials, under a partial vacuum in order toeliminate air bubbles, into a homogeneous fluid mass that forms the final sheet. Theextruder produces a sheet approximately 0.1 to 5 mm in thickness and up to about10 m wide.Multi-layer, unreinforced or reinforced geomembranes are somewhat differentand more complicated to manufacture. As the name indicates, this type ofgeomembrane consists of two or more laminated plastic sheet layers that can bereinforced by placing a fabric scrim between the plastic layers.Spread coating is the third and most recent production method forgeomembranes. This technique uses geotextiles in combination with the moltenpolymer. Usually a needle-punched nonwoven geotextile is used as a substratum uponwhich the molten raw materials are placed and spread into its final thickness.1.1.4 GeonetsGeonets are the smallest and the most recent segment of the geosyntheticproduction range. Geonets have a netlike configuration similar to that of geogrids, butserve a different function, and are typically made of polyethylene only. Themanufacture of geonets begins with blending of the raw materials, which are fed into anextruder where the polymeric material is melted. It then passes into a counter rotatingformer which produces ribs at acute angles to one another; pressure is used to push theextrudate forward over a steel spreading mandrel which opens the ribs in a relativelylarge diamond-shaped apertures, typically 12 mm long by 8 mm wide, and 5 to 7 mmthick.Chapter 1 - Introduction^ 51.1.5 GeocompositesGeocomposites are, as the name suggests, a combination of severalgeosynthetics. The concept is to use the best features of the other geosynthetics and,by combining the different attributes for two or more materials, achieve an improvedperformance for specific applications.1.2 FUNCTIONS AND APPLICATIONS OF GEOSYNTHETICSSince the early 1950's when the first geotextiles were introduced(van Zanten, 1986), the use of geosynthetics has been playing an increasingly importantrole in various engineering applications. There are numerous applications wheregeosynthetics are used, and their main functions can be divided into five categories:• separation;• reinforcement;• filtration;• drainage; and• low permeability barrier.These functions will be discussed subsequently, with emphasis placed on the useof geotextiles and geogrids in unpaved road construction.1.2.1 SeparationGeosynthetics are commonly used to satisfy more than one function. When ageosynthetic is used as a separator, it must prevent the intermixing of particles fromtwo soil layers with different properties, for example between a course base aggregate,or ballast, and a softer subgrade soil.Chapter 1 - Introduction^ 6Unpaved roads fall within this application. The prime function of geosyntheticsin unpaved road construction is often considered to be separation, the secondaryfunctions being reinforcement and filtration. However, this situation is very dependenton the local conditions and might be true for some cases but over all, the reinforcementfunction is the primary function that geosynthetics serve today.When designing for separation there are various material properties that must betaken into account to determine quantitatively which geosynthetic is suitable for aspecific application, such as tensile strength, and resistance to tear, burst and impact.There is a common misunderstanding between separation and filtration, and thesetwo different but related functions are very often confused or considered as being thesame. Separation might at first sight appear synonymous with filtration but the maindifference is that unlike filtration the separation role is performed independently of thepresence of water. Numerous applications can be found but only a few of them will bediscussed and described herein. The products in the geosynthetic group do not allperform this function, which is generally referred to as the first identified function ofgeotextiles even though it is often underrated.1.2.2 ReinforcementSynthetic reinforcement is increasingly used for temporary and permanent earthstructures. There are three main areas of application: slopes and embankments;foundations; and retaining walls. While specific applications are numerous, only one ofthem will be discussed in some detail, which is the reinforcement application inunpaved road construction.Chapter 1 - Introduction^ 7The reinforcing function of geosynthetics, laid between the subgrade soil and thebase course aggregate can be categorized by: subgrade and base course restraint;lateral restraint of base course; and tensioned membrane effect.- Subgrade and base course restraintWhen traffic load is applied to a base course layer tension forces are developed atthe base and the cohesionless material tends to move apart, allowing intrusion of thesubgrade and mixing of the two materials. Geosynthetics provide a tensile reinforce-ment, and when placed at the subgrade-base course interface, base course aggregatemovement is restrained. The bearing capacity of the subgrade layer is also greatlyreduced by outward acting shear stresses at the subgrade surface. The presence of ageosynthetic prevents these stresses from being transmitted to the subgrade, allowingthe full bearing capacity of the subgrade to be mobilized. The presence of thegeosynthetic has been found to prevent local shear failure in a soft subgrade soil andincrease the bearing capacity towards a general shear failure.- Lateral restraintThe lateral restraint provided by the presence of the geosynthetic increases theload distribution effectiveness of the base course layer resulting in decreasingcompressive stresses on the subgrade.- Membrane effectThe tensioned membrane effect is mobilized when the subgrade is deformed byrutting and as the subgrade deforms the geosynthetic stretches and tensile stresses aredeveloped. The vertical component of the tension in the geosynthetic acts to reducethe load transmitted to the subgrade.Chapter 1 - Introduction^ 81.2.3 FiltrationFiltration is defined as the promotion of a natural filter in the adjacent soil: wateror liquid can pass through the plane of the fabric while the soil is retained. The use ofgeotextile filters can be subdivided into three categories (John, 1987), based on theflow conditions. These are listed below in ascending order of severity:• fairly steady unidirectional flow;• reversing flow with a moderate cycle time; and• reversing flow with a very short cycle time.Examples of applications corresponding to these three flow conditions are respectively:• land drainage filters;• river and coastal defense filters; and• anti-pumping filters beneath ballast under railroads.Other applications include filtration beneath stone base for unpaved roads andairfields, around a perforated under-drain pipe, as a silt fence and/or curtain, and as afilter beneath stone rip-rap. Three main factors that must be taken into account inorder to properly design a filter are; adequate permeability, soil retention, and longterm compatibility.The permeability criterion is one of the main factors in selecting an appropriatefilter. However, the coefficient of permeability is difficult to determine accurately for ageotextile. Thick geotextiles are compressible, hence changes in applied stress causeschanges in permeability. For these reasons the thickness is taken into account and anew term, permittivity, is defined.Chapter 1 - Introduction^ 9The requirements of a high permeability and good soil retention are contradictorysince the cross-plane water flow would be least restricted if the openings were large,whereas small openings would be most effective for a good soil retention.Consequently an upper and lower bound value of characteristic opening size is oftenspecified in design.Soil to fabric compatibility is one of the questions that have to be answered. Aunidirectional flow promotes the formation of a graded filter in the adjacent soil whenthe pore size openings of a geotextile are compatible with the distribution of the soilparticle sizes. However, if the geotextile openings are too large then piping will occur,either as an internal erosion or as soil suffusion (John, 1987), depending on the soilparticle size grading.In contrast, if the pores are small a graded filter may not form since insufficientinitial wash-through of fine soil particles occurs. Blocking is likely to occur if thepores of the geotextile are too small, or even blinding, which is a special form of severeblocking. When soil particles sizes are very close to that of geotextile pore openingssome of the soil particles are likely to enter the pores and cause clogging.These factors are all important when designing for filtration and have only beenbriefly described. Filter design for various applications, soil conditions, fabric types,and flow regime is not an easy task since there are few good experimental and fielddata to support relationships for flow conditions other than unidirectional.1.2.4 DrainageDrainage or fluid transmission is the fourth basic function of geosynthetics. Theterm drainage refers to the flow of water or liquid within the plane of the geosynthetic.Geotextiles, geonets and geocomposites are materials frequently used to provide arelatively high flow capacity path within a soil structure along which water can flowChapter 1 - Introduction^ 10preserving good soil retention and a long term soil-to-geosynthetic compatibility.Some selected applications are listed below:• chimney drain and drainage gallery in earth dams;• wick drains;• drainage behind a retaining wall;• pore water dissipation in earth fills;• leachate drainage of landfill side slopes and above landfill liners;• drainage of frost susceptible soils;• drainage beneath a surcharge fill and building foundations; and• pavement edge drain.When designing with geosynthetics the two general flow capacity categories(Koerner, 1990), are gravity flow and pressure flow. Gravity flow is governed by theslope while pressure flow is independent of orientation but occurs generally fromhigher pressure to locations of lower pressure.Physical properties such as thickness, mechanical properties such as tensilecompressive and shear strength, hydraulic properties, endurance properties andenvironmental properties all have to be considered in drainage design, where some ofthem are not as critical as others depending on the application.1.2.4 Low Permeability BarrierA low permeability barrier is used to prevent or minimize liquid movement fromone point to another. Formerly this was done by using clay or other poorly drainingsoils, which dates back many centuries. Asphalt and various cements have been usedsince early this century, and more recently concrete and synthetic membranes havebeen developed.Chapter 1 - Introduction^ 11Geomembranes have a very low permeability to both gases and fluids, and areused extensively as low permeability barriers. The potential applications ofgeomembranes are: sealing against fluid percolation; and as buffers against pollutants.The function of the geomembrane in case of the former application is to form a barrierbetween water or other liquid and the surroundings, and to keep liquid flow throughthe membrane to a minimum. The second function involves attenuation ofcontaminants during migration. Geomembrane properties and laboratory tests forgeomembrane characterization according to the Canadian Foundation EngineeringManual are as follows:• identification properties; such as:thickness;density;- mass per unit area;melt indexhardness;- carbon black content; andcarbon dispersion;• performance tests; e.g.:tear;- tensile properties;bursting;- puncture;- friction;- creep;- water and vapor permeability;Chapter 1 - Introduction^ 12• performance tests on joints; e.g.:- shear;- peal; and- non destructive tests;• Durability; e.g.:chemical resistance;- biological resistance;- stress cracking;- abrasion;- thermal aging; and- light exposure.1.3 SCOPE OF RESEARCHWhile significant work has examined the mechanisms of reinforcement andstabilization at model scale in the laboratory, few well-controlled field tests have beenperformed and rigorously evaluated. This research project involves construction of an80 m long unpaved road that was loaded using a special pavement test vehicle with astandard axle load. The project is unusual because the test road is not part of anexisting highway and the only traffic loading is the pavement test vehicle. Aconsortium of three manufacturers of geosynthetics, and the B.C. Ministry ofTransportation and Highways, are supporting the construction and trafficking test.Four sections of the road include a geosynthetic, and one is unreinforced. Simulatedtraffic loading is intended to evaluate the benefits of using geosynthetics in unpavedroad construction.Chapter 1 - Introduction^ 131.4 RESEARCH OBJECTIVESThe main objectives of this study are to:• quantify the improvement of a geosynthetic reinforced sections overan unreinforced section based on measurements of rut depths versusnumber of passes, base course and subgrade deformation profiles,and geosynthetic strain measurements;• compare relative performance of different geosynthetics;• identify the relative importance of the basic functions:separation;reinforcement; andfiltration;• critically evaluate design methods1.5 OUTLINE OF THESISThis chapter has provided an introduction to the fundamentals of geosynthetics.In chapter two, the literature related to this research project is reviewed, with specialemphasis placed on bearing capacity, model studies, analytical studies and other fieldstudies. Design procedures are also reviewed. The test site is described in the thirdChapter 1 - Introduction^ 14chapter. A short geological history of the Lower Mainland is given as well as surficialgeology of Lulu Island. The location and description of the test site are described andsubgrade soil properties reported. Chapter four deals with material properties of thebase course and the properties of the four different geosynthetics used in the testsections. It also describes the five test sections, with reference to construction designand instrumentation. A short description is given of the loading vehicle used in the trialand the trafficking schedule.A comprehensive summary of the field measurements is given in chapter five.Since huge amount of data was collected during the tests, there is no possibility ofincluding all this \"raw\" data in the thesis. Chapter six contains an analysis andinterpretation of the results. A comparison is made between the test sections and theresults are compared to design methods and other available field trials. The lastchapter contains recommendations and the conclusions from this research.Chapter 2Literature Review2.1 INTRODUCTIONGeotextiles and geogrids have been used successfully in unpaved roadconstruction over soft ground for many years. The geotextiles often fulfill more thanone of the basic functions of separation, reinforcement, filtration and drainage in theirapplications in unpaved roads. Geogrids, however, serve mainly as reinforcement andin some cases as separators.Over the past two decades a considerable amount of research on geosyntheticreinforcement of layered soil systems has been carried out internationally. Thesestudies have been in the form of field, model and analytical studies, and have supportedthe development of design procedures for unpaved road construction over soft ground.Although most of the design procedures are similar in their approach, differentassumptions are made with respect to vehicle traffic, bearing capacity formulation, the15Chapter Two - Literature Review^ 16effect of anchorage, and the mechanism of stabilization, particularly the relativeimportance of a tensioned-membrane effect.2.2 FUNDAMENTALS OF DESIGN APPROACHESAlthough geosynthetics have been used in road construction for over twodecades, the design is based largely on empirical rules. Theoretical work has, however,grown to make a significant contribution based on interpretation of field observations,small and large scale laboratory models, analytical work and sophisticated finiteelement studies. The current understanding of geosynthetic-soil interaction and theprocedures used in design are described below.2.2.1 Bearing CapacityThe ultimate load that a soil body can support is termed its bearing capacity, andis dependent not only on the mechanical properties of the soil but also on the shape,size and the location of the loaded area.The first rational approach to estimate the bearing capacity of soil was providedby Terzaghi (Terzaghi, 1943), who reasoned that if the weight of the soil, the effects ofthe surcharge, and the strength parameters of the soil could be assessed on the basis ofa conservative mechanism, then the sum of these factors would in turn provide aconservative estimate. In his bearing capacity formulation he differentiates betweengeneral shear failure, characterized by recognizable failure planes extending from theedge of the loaded area, and local shear failure, where plastic flow causes largesettlement under the footing without noticeable bulging at the surface. He alsodistinguished between smooth and rough footings.14 ^ BN0 14—B/2-111+111--B/2 —0114^qd^qdic/4-1-40/2111 11/4-0/2^ill4)ft/2—702+ q 0q0--1114-4)/2^M^X/4-40/2r BNaChapter Two - Literature Review^ 17The conditions for general shear failure of a shallow continuous footing, seeFig. 2.1, show the zones assumed for a smooth and rough footing. The shearingresistance of the soil located above the level of the base of the footing can be neglectedif the condition of a shallow footing is satisfied, that is Df is less than B, for which soilcan be replaced by a surcharge (to equal to Dfy per unit of area, where y is the unitweight of the soil and Df is the distance between the surface of the ground and the baseof the footing.Smooth footing^Rough footingFigure 2.1 Failure of soil beneath continuous smooth and rough footing.The zone of plastic equilibrium for the smooth footing is defined by the areaACDE, and by the area A1CD'E' for the rough footing. The difference is the assumedboundary between the elastic zone I and the zones of radial shear II: for the smoothfooting it is inclined at an angle of 7c/4+4)/2, whereas for the rough footing it is inclinedat an angle 4). Zone I for both cases is assumed to remain in the elastic state because offriction and adhesion between the soil and the base of the footing, and acts as a part ofChapter Two - Literature Review^ 18the displaced footing. This downward movement produces outward lateral forces onboth sides of the wedge and a passive Rankine state is developed in zones M r and M.Terzaghi's classical bearing capacity formula is based on earlier solutions byPrandtl and Reissner, and the approximate value of the ultimate bearing capacity isgiven by the equationqu = cN + yD f Ng yBN7^ (2.1)in which No Nq and N are dimensionless bearing capacity factors related to thecohesion (c), surcharge and weight of the soil, respectively, and only dependent on theangle of internal friction. The values of Nc and Nq are determined by the followingequations derived by Terzaghi (Terzaghi, 1943) front those published by Prandtl and byReissner for continuous footings with a rough base:Is.rc cot 4 a; 1]2 cos 2 (45°+ 0/2)(2.2)N =^452 (4 q^2 cos °4- 012) (2.3)wheregivingNir-012) tan #a = e (9N = cot 4N q —1)(2.4)(2.5)Currently the most widely used values for the factor N y are those obtained by Hansenand Meyerhof (Craig, 1992), represented by the following approximations:N7 =1.80(Nq —1)tan 0 (Hansen)^(2.6)A T = (Nq — 1) tan(1. 4 0) (Meyerhof)^(2.7)Chapter Two - Literature Review^ 19The boundaries that CD and CD' define (Fig. 2.1) are generally log spirals. If however= 0° the spirals become arcs of circles and the corresponding values of thedimensionless bearing capacity factors Nc, Nq and Ny become:N,=in+1= 5.7, Ng = 1, and Nr = 0Hence, for a rough footing at the ground surface the value of the bearing capacity perunit area is:qz, = (1 n +1)e = 5.7c^(2.8)The equations above refer to footings with a rough base where the boundarybetween zones Ir and IIr is inclined at an angle 4), see Fig. 2.1. If the frictionalresistance of the footing reduces, the angle between I and II increases. Assuming anangle 13 for the case of a footing having less resistance to sliding than the roughfooting, but not perfectly frictionless, where 4) < 13 < 45°14/2, the equations become:Arc = tan/3+ sc°sin 0°3:0:6; [a 02 (1 + sin 0) —1]^(2.9)and(2.10)Nq = coss(fr-, a 92 tan(45° + —2'6 )wherea = e r+ 0/24 tan # (2.11)Assuming a perfectly frictionless base, i.e. f3 = 45°+4)/2:N = cot 4a20 tang (45° + 2) —11^(2.12)andNq = aB tan2 (45° +^(2.13)whereIn-tana9 = e 2 (2.14)Chapter Two - Literature Review^ 20If the footing is assumed perfectly smooth and if 4) = 0° the bearing capacityfactors Nc, Nq and N become:Ng = 7r+ 2 = 5.14, Ng = 1, and Nr = 0Hence, for a smooth footing at the ground surface the value of the bearing capacity perunit area is:qi, =(7r+2)c = 5.14c^(2.15)The above procedure is for the case of general shear failure. For the case of localshear failure the values of c and 4) must be replaced by c' and 4) 1 in the equations abovein order to calculate the corresponding values N' c, N'q and NI1, . This is done by using(Terzaghi, 1943)2C = -3 C (2.16)andtan 0. = 3 tan^ (2.17)in the above equations, and the bearing capacity is then given by:qu =^+ rD f N + yBN,^(2.18)Hence, for a smooth footing at the ground surface the value of the bearing capacity inthe case of local shear failure is:qu. =icNg = 3qu = 3(,r+ 2)c = 3.43c^(2.19)It should be noted that Terzaghi's bearing capacity derivation above is basedupon limit equilibrium conditions and therefore the resultant bearing capacities are aChapter Two - Literature Review^ 21measure of the ultimate failure pressure that can be applied to a footing. Furthermore,the preceding discussion refers only to continuous footings, and for shapes other thanstrip footing the bearing capacity equation has to be modified. Based on experimentalresults, the following semi-empirical equations are used (Terzaghi and Peck, 1967) fora circular footing of radius R:quR =1.2cN, + yDfNq + 0.6 yRArr^(2.20)and a square footing, BxB:quc = 1. 2cN, + yDfNq + O. 4 yBNr^(2.21)Hence, the bearing capacity of a footing at the ground surface, for (1) = 0, is given by:q„R = quc =1.2(g+ 2)c = 6.17c^(2.22)To account for the geometric-loading and depth-influence factors, Skempton proposedthe following expression for values of Df113_2.5 in the case of a rectangular footing oflength L:qu = (g+ 4(1+ 0.2 ',6-`)(1+ 0 2 /lc• L (2.23)A width to length ratio of unity gives the same value of bearing capacity as eqn. (2.22).For two layer systems, several methods have been developed for calculation ofthe ultimate bearing capacity. Purushothamaraj et al.(1973) used Prandtl-Terzaghi'smechanism for a rough shallow footing, assuming that both soil layers arehomogeneous and isotropic. By applying the kinematics consideration of the Drucker-Prager (1952) second theorem, they developed a theoretical procedure to estimate thebearing capacity factors. Meyerhof (1974) presented analyses of different modes ofsoil failure for a dense sand overlying a clay, and compared the results to model testsChapter Two - Literature Review^ 22and field observations. He showed that the influence of the sand layer thicknessbeneath the footing was governed primarily by the bearing capacity ratio of the twolayers, the friction angle of the sand, and the shape and location of the footing.In the case of reinforced unpaved roads several variations of the bearing capacityequation have been proposed, where the placement of a geosynthetic on the softsubgrade is believed (Bakker, 1977; Barenberg et al., 1978; Giroud and Noiray, 1981;Sellmejer et al., 1982) to promote a general shear failure where otherwise a localfailure would occur. Although there is some difference of opinion on the relativeimportance of the physical actions of the geosynthetic, the increase in bearing capacityseems to be well accepted.2.2.2 Membrane EffectThe concept of a tensioned membrane effect due to the geosynthetic whichdevelops as the road undergoes deformations, is illustrated schematically in Fig. 2.2..In order to develop any benefit from a membrane action, the geosynthetic must beanchored effectively outside the loaded area and significant vertical deformations mustoccur.Figure 2.2 Membrane action of geosynthetic in unpaved road.Chapter Two - Literature Review^ 23Stress-strain characteristics of the geosynthetic are also important when consideringthe membrane action. The higher the modulus of the geosynthetic, the less verticaldeformation is required to develop the equivalent support from an inclusion of lowermodulus.The membrane effect is attributed to the in-plane tensile stress in the geosyntheticinducing stresses perpendicular to the plane of the geosynthetic: the imposed stresseson the subgrade in the concave up-section of the membrane beneath the wheels isreduced and the stress in the adjacent concave-down section outside the wheel path isincreased.Several tension membrane models are available in the literature (Barenberg,1980; Giroud and Noiray, 1981; Raumann, 1982; and Sellmeijer et al., 1982). Theyvary in complexity, but result in a similar contribution to the predicted bearingcapacity. In order to clarify some of the assumptions which are made, a simple modeldeveloped by Gourc et al. (1982) will be described briefly, see Fig. 2.3.14- B'Figure 2.3 Membrane effect in unpaved roads with geosynthetics - basic parameters.Chapter Two - Literature Review^ 24The deformed shape of the geosynthetic was approximated by three circular arcsegments of equal radius, for whichand2 e = R — cos, e = 2R(1 — cos 0) (2.24)B' = 2R sin 9^ (2.25)The geometric relationship between the deflection of the geosynthetic (e) and thesettlement through diameter B', is therefore (Fig. 2.3):e 1 — cos 0B'^sin 0(2.26)The strain (6) in the geotextile assuming perfect anchorage and a constant tension overthe width of the membrane is equal to the elongation due to the deformation over theinitial length B', is given by:Al _ /2 —^ = 2 OR— 2R sin 0 0— sin 0e = 4^4^2R sin 0^sin 0(2.27)The tensile force per unit width in the geosynthetic, assuming a linear stress-strainrelationship, is therefore:T = Ee= E 0— sin 0sin 0(2.28)The component of this tension which acts vertically upward, reducing the imposedstress on the subgrade in the concave-up section of the membrane underneath thevehicle wheel, is:AqB' = 2T sin 9= 2E(9— sin 0)^(2.29)Chapter Two - Literature Review^ 25The reduction of applied stress is therefore strictly dependent on the rut depthand the modulus of the geosynthetic. It should be noted that the above procedure isonly one of several methods which all are approximations to what really takes place insitu. Some shortcomings in the assumptions which have been established from testingare an underestimation of lateral slippage between the geosynthetic and the subgradesoil and a different tensile stress distribution in the deformed geosynthetic(Jarrett, 1986).2.2.3 Effect of AnchorageMany design methods consider the role of the tensioned membrane effectimportant to the increased bearing capacity of reinforced unpaved roads, and placeemphasis on the need for anchorage of the geosynthetic outside the loaded area tomobilize it. Several papers have been published which discuss geosynthetic-soilinteraction and anchorage (Holtz, 1977; Myles, 1982; Jarrett and Bathurst, 1985;Douglas and Kelly, 1986; Bourdeau and Holtz, 1988; and Dembicki, 1991.Unanchored and anchored tests conducted by Douglas et al.(1985) and Douglas andKelly (1986) indicated no significant variation with different anchorage details for thegeosynthetic used in their laboratory study. Similar conclusions were also reported byde Garbled and Javor (1986). Douglas and Kelly (1986) also observed from their teststhat the overall stiffness of the model road did not seem to change significantly with adifferent tensile modulus of geosynthetic, for the range examined in their study.Dembicki (1992), who conducted model tests to examine the friction of geotextiles andthe effect of anchorage on the bearing capacity of the subgrade, concluded the lengthof anchorage section is of a great importanceA difference in stiffness for unreinforced and reinforced tests with single and dualfooting was observed in a small scale model test conducted by Love et al.(1987).Chapter Two - Literature Review^ 26Where for the single footing no significant difference was observed between theunreinforced and the reinforced case, whereas when the load was applied by using dualfooting a considerable difference was observed as well as increase in stiffness betweenthe two loading patterns, affirming that the loading pattern is of importance whencomparing the performance of reinforced and unreinforced unpaved roads.The observations of Douglas and Kelly (1986) did not meet with universalagreement and an interesting debate on this matter has ensued, see Douglas andKelly (1986), Bourdeau and Holtz (1988), Douglas (1990), and Bourdeau andHoltz (1992). The main conflict arises from their interpretation of the results, sincethey show a difference in behavior which is less that might be expected for ageosynthetic-soil system.In summary, the mechanism of geosynthetic-soil interaction in unpaved roadsappears to be complex and is not fully understood. The effect of anchorage, andconsequently the contribution of a tensioned membrane effect, is not well-defined.Since the basic function that a geosynthetic serves in an unpaved road varies with thetype of vehicle loading, subgrade strength and base course material properties, it wouldseem inappropriate to consider only one conceptual model when interpreting results ofstudies.An interesting approach to the design of unpaved road using geosynthetics,which might explain some of the contrariety existing in the literature, has beendeveloped, Milligan et. al (1989). Their method which is based on extensive researchprogram consisting of model test under monotonic loading (Love, 1984), under cyclicloading (Fannin, 1987), and finite element analysis (Burd, 1986), disregards themembrane effect and therefore the effect of anchorage as well for small deformations.Chapter Two - Literature Review^ 27Instead it is postulated that the outward acting shear stresses are picked up by thegeosynthetic and only vertical stresses are transmitted to the subgrade below. Theirprocedure will be described in some detail in section 2.4.2.3 UNPAVED ROAD CONSTRUCTIONOver the past fifteen years a number of methods for designing with geosyntheticsin unpaved roads have been published. The design approaches are typically based onfield studies, model studies or analytical studies, or in some cases a combination of thethree.2.3.1 Field StudiesFew well-controlled field trials have been performed to examine the influence ofgeosynthetics in unpaved roads. Early work, to which much reference made in morerecent design methods, is that of the U.S. Army Waterways Experiment Station(Webster and Watkins, 1977; Webster and Alford, 1978). Webster and Watkins(1977) report details of a field trial involving seven test sections: one was reinforcedwith a nonwoven spun-bonded needle punched polyester geotextile, one with aneoprene-coated one-ply woven nylon membrane, and one was left unreinforced to actas a control section. The remaining sections were reinforced with materials other thangeosynthetics. The test sections were 9.1 m long, and 3.7 m wide; They comprised a35 cm thick crushed stone base course placed on a 61 cm thick clay subgrade ofundrained shear strength 21 - 30 kPa in the upper 25 cm and 30 - 69 kPa below.Vehicle loading was applied using a 44.5 kN tandem axle dump truck, but analyzed interms of an equivalent 80.1 kN single axle, dual wheel load. The performance of theChapter Two - Literature Review^ 28sections was evaluated from visual observations, photographs, and deformations ofcross-sections recorded at intervals throughout the trafficking period. The sectionswere trafficked up to an average rut depth of 28 cm, which can be considered asfailure. Results showed the performance of the reinforced sections was significantlybetter than the unreinforced section. The unreinforced section accepted 200 passes,the non-woven geotextile section 2500 passes, and the woven membrane 37000 passes.This investigation was probably the first clear affirmation that usinggeosynthetics in construction of unpaved roads on soft ground leads to a significantimprovement in traffickability.Other field trials are reported (Potter and Currer, 1981; Sowers et al., 1982;Ramalho-Ortiago and Palmeira, 1982; Ruddock et al., 1982; Delmas et al., 1986;de Gardiel and Javor, 1986; Itoh et al., 1990; Austin and Coleman, 1993) which vary intheir construction procedure, base course and subgrade soil properties, and imposedtrafficking. Yet they all have one thing in common: the reinforced sections performsignificantly better than unreinforced sections.Besides reduced surface deformation due to the presence of the geotextile at thebase course-subgrade interface, Potter and Currer (1981) also observed a considerablereduction in transient horizontal strains in the subgrade. The presence of the fabric didhowever not influence the amount of vertical stress and strains, and similarobservations were reported by Ruddock et al. (1982). Sowers et al. (1982) reportedtheir results for both light and heavy loads. They observed in the case of a light loadthat the presence of the geotextile helped to maintain a load spreading action in thebase course by preventing intrusion of the subgrade into the base course material aswell as providing restraining effects of the base course layer. For heavier loads, andincreased rut depths, a tensioned-membrane effect was observed; when the subgradewas re-graded by filling in the ruts after the fabric had developed substantialChapter Two - Literature Review^ 29deformations, a further significant improvement in traffickability was observed that wasattributed to the tensioned-membrane effects.The most interesting observation in the field test described by Ramalho-Outrageand Palmeira (1982) is the effect of different anchorage details on the performance ofan unpaved road. However, it must be noted that the base course was not a typicalrecommended base course material for unpaved roads, it consisted of a residual clayeysoil and the thickness of the base course was therefore approximately 1 m instead of 30cm thick if a good material would have been used. Field measurements showed 10 to24% volumetric savings of the fill material when geotextiles where used where thedifference in savings is attributed to the different anchorage details.Delmas et al. (1986) performed a full scale test using four different geotextiles,two nonwoven and two woven geotextiles, of different tensile stiffness. The resultsshowed a considerable difference in rut development with number of passes for thedifferent fabrics, with the stiffer woven materials giving the best performance. Theyalso observed a better performance for a given fabric if the anchorage length outsidethe wheel path was longer. De Garbled and Javor (1986) also observed a betterperformance from reinforced sections in their field test, using three different geotextilesof tensile strength in the range 35 kN/m (non-woven) to 480 kN/m (woven).However, they did not detect any significant difference in rut development between thethree different geotextiles; this might be due to heterogeneity in the subgrade soil.Similar observations were reported by Austin and Coleman (1993).From these full scale field trials it is evident that the inclusion of a geosyntheticdoes improve the performance of an unpaved road over soft ground. It is, however,not clear what the main function of the geosynthetic is, and further well-documentedfield trials are very important to the provision of good quality data for calibration ofapproaches used in design.Chapter Two - Literature Review^ 302.3.2 Model StudiesNumerous model tests, both at large and small scale, have been performed toanalyze the bearing capacity and deformation behavior of a base course layer over asoft subgrade, with a geosynthetic inclusion, e.g.,(Jessberger, 1977; SOrlie, 1977;Jarrett et al., 1977; Bell et al., 1977; Barvashov et al., 1977; Andersson, 1977; Robnettet al., 1980; Barksdale et al., 1982; Robnett and Lai, 1982; Gourc et al.,1982; Kinney,1982; Raumann, 1982; Milligan and Love, 1985; Jarrett and Bathurst, 1985; Laier andBrau, 1986; Douglas and Kelly, 1986; Bauer and Preissner, 1986; De Garidel andJavor, 1986; Resl and Werner, 1986; Jarrett, 1986; Milligan et al., 1986;Love et al., 1987; Fannin, 1987; Floss et al., 1990; Houlsby and Jewell, 1990; Burd andBrocklehurst, 1990; Guido et al., 1991; Alenowicz and Dembicki, 1991; Dembicki,1991; Douglas and Valsangkar, 1992). The review of the literature is however quiteconfusing since conflicting observations are reported in some of these studies andindicate everything from inferior to superior performance of unpaved roadincorporating geosynthetics compared to roads without any reinforcement.Due to the amount of available reported model studies a review herein would beto comprehensive and difficult to choose between reported studies in order tocondense the review.2.3.3 Analytical StudiesThe tensioned membrane effect which is mobilized by deformations of thegeosynthetic in an unpaved road over soft subgrade is considered as being of greatimportance in most of the available design methods. Some of the published methodsfor calculating the membrane effect are summarized briefly below.The first attempt at describing the behavior of the tensioned membrane was madeby Nieuwenhuis (1977) and Bakker (1977), who both observed the importantChapter Two - Literature Review^ 31interaction between development of deformations in the subgrade and the tensile forcesin the fabric. Nieuwenhuis (1977) assumed a linear elastic behavior of the membrane,and described the stress-strain characteristics of the subgrade by employing acoefficient of subgrade reaction or a spring constant; vertical equilibrium wasaddressed using Boussinesq theory for calculation of the load transferred through thebase course layer. Bakker (1977) considered the ultimate bearing capacity of thesubgrade, which was modified for the case with a geotextile by assuming a bilineardeformed shape of the geotextile under the loaded area, and a linear elastic behavior inthe membrane.Giroud and Noiray (1981) followed a similar approach, but assumed thedeformed shape as parabolic. Sowers et al. (1982) also assumed a parabolic deformedshape for the geotextile, but instead of assuming a linear stress-strain relationship asothers had done, they employed a strain-dependent tension in the fabric; strains werecalculated from an empirical formula, and the tension established from approximatestress-strain curves.Bourdeau et al. (1982) followed a similar approach to that of Nieuwenhuis(1977) but instead of using a Boussinesq solution for the semi-infinite half space forcomputing the load distribution through the base course layer, Bourdeau et al. (1982)used a probabilistic concept for the vertical stress diffusion in a particulate media.Assuming the subgrade deformation characteristics as a Winkler media the governingdifferential equation for vertical equilibrium is solved.Sellmeijer et al. (1982) developed a method for calculating the tensionedmembrane effect from a theoretical consideration that satisfies both subgrade andgeotextile equilibrium. The equilibrium equations are then solved simultaneouslyassuming a bilinear elastoplastic behavior in the subgrade, and adequate anchorage ofthe geotextile. Sellmejer et al. (1982) will be discussed in more detail subsequently.Chapter Two - Literature Review^ 322.4 REVIEW OF DESIGN PROCEDURESSeveral design methods for unpaved roads incorporating geosynthetics have beenpublished over the last two decades. Many have been developed for specificcommercial products, but others are intended to be generic. They may be grouped intothree broad classes: the first is empirical; the second is classified as semi-theoretical,where field experience and experimental data are used in association with availabletheory; the third approach is purely analytical, and includes the finite element method.2.4.1 Empirical Design ApproachesFew empirical design methods are available in the literature, and only one thatcan be classified as being strictly empirical was found in the literature review. Thismethod, described by Jaecklin (1986), was developed using a sophisticated regressionanalysis on a large data base of case histories, and a survey of European. It takes intoaccount traffic, rut depth, subgrade strength, base course aggregate properties andthickness, and a geotextile factor which is dependent on the mobilized strain in thefabric. These design factors are defined as follows:V - factor : Traffic or vehicle loadV = 0.5 very light vehicles and carsV = 1^light vehicles^maximum 10 trucks per day.V = 2^medium^10 - 50 trucks per dayV - 2.5^heavy^50 - 100 trucks per dayV = 3^very heavy^more than 100 trucks per dayChapter Two - Literature Review^ 33Cars and light vehicles are considered to weigh less than 3500 kg (34.3 kN) and truckhas a total weight of 36 tons (353.2 kN) on 4 axles or 9 tons (88.3 kN) per axle.R - factor : Rut depth^K - factor : Base course aggregateR = 3-4 rut depth up to 3 cm^K = 0.5 crushed rockR = 5^rut depth up to 5 cm K = 1^clean gravelR = 10^rut depth up to 10 cm^K = 2^silty gravelR = 15^rut depth up to 15 cmU - factor : Subgrade strengthU = 1 (firm)U = 2 (soft)U = 3 (very soft)CBR of 5-10 %, or Su = 150-300 kPaCBR of 2-5 %, or Su = 60-150 kPaCBR of 1-2 %, or Su = 10-60 kPaD - factor : Base course thickness The D factor is related to the traffic and subgrade strength which determines theminimum required base course thickness.Traffic typeMinimum base course thickness(cm)U= 1 U = 2 U = 3V = 1 30 35 40V = 2 35 40 45V = 3 40 45 50The D factor is then determined based on the minimum base course thicknessrequirements: D = 3, 4 and 5 for base course thickness of 30 - 40 cm, 40 - 50 cm and50 - 80 cm, respectively.Chapter Two - Literature Review^ 34The minimum failure force, F 1 in lcN/m, is then calculated according to the empiricalformula shown below:F • =G[5+2.1x(V +0.8) 0.8 x 4 X 0.9U1-2 X^X ill-F( YK — 1) 2^(2.30)31?\"Where G is assumed to be 1 if failure strain is higher than required minimum strain andif the failure force is larger than the required minimum too. Otherwise, if failure strainis not as high as determined the G - value is calculated as followserequiredG= reffectiveer(2.31)The empirical equation for calculating the required minimum strain in the geotextile atfailure is:1.4 x logR 4” +10+10 xlogD)= v^°' x K\"The actual failure strain is then determined by:nc )c„ e0.2457Pr^•(2.32)(2.33)where1 P =14 E )0.2457f^— 16.913(2.34)The above formulas can then be used to produce various design charts, such as failurestrain and failure force versus base course thickness for various traffic loading, rutdepths, base course types and subgrade or a more common chart with required basecourse thickness versus subgrade strength for different rut depths, loading and failureforce.Chapter Two - Literature Review^ 352.4.2 Semi-Theoretical Design MethodsThe most rational basis for a design method is indisputably a combination of fieldexperience, experimental data, and theory. There are many design methods that aresemi-theoretical, but they differ in how they combine theory and experience and towhat extend different design factors are considered and the results interpreted.Barenberg et al. (1975) and Bender and Barenberg (1978) developed a simpledesign procedure based on a laboratory testing program using large scale threedimensional and small scale two dimensional model systems of a nonwoven geotextilereinforced aggregate-subgrade system. Based on their results they concluded thatfailure in the unreinforced system occurred when the ratio of applied stress to theundrained shear strength of the subgrade was about 3.3 in the unreinforced system andabout 6 in the reinforced system. These conclusions compare well to Terzaghi'sbearing capacity factors for general and local shear failure, respectively. Fordetermining the base course thickness when a geotextile is used in construction, thevertical stress is calculated using Boussinesq theory, where the allowable stress on thesubgrade is not to exceed six times the undrained shear strength of the subgrade. Inthe unreinforced system, the same procedure is used but a bearing capacity factor of3.3 rather than 6 is used. It should be noted that in the research program only one typeof geotextile was used, Mirafi 140, and therefore caution should be used when applyingthe results to other fabrics.Stewart et al. (1977), modified the recommendations of Barenberg aftercompleting a full scale field trial where seven different geotextiles where used forreinforcement in an unpaved road over a soft subgrade. It was concluded that deeprutting, which was defined as being about 10 cm, would occur at stress levels of 3.3c uwithout a fabric and at 6.0cu with fabric. The values are the same as those reported byBarenberg earlier, but Stewart et al. (1977) also established that very little rutting (lessChapter Two - Literature Review^ 36than 5 cm) would occur if stress levels in the subgrade did not exceed 2.8cu in theunreinforced system at high traffic volume and 5.0cu in the reinforced case.Based on the tensioned membrane model developed by Kinney andBarenberg (1978,1982), Barenberg (1980) modified his original design approach inorder to include the tensioned membrane effects. The applied stress on the subgrade iscalculated according to Boussinesq theory and the base course thickness (h) thencalculated by setting the applied stress equal to the sum of the permissible stress andthe differential normal stress across the fabric. The vertical stress underneath thecenter line of the applied load is calculated assuming circular contact area using thefollowing equation.P^h3Crz = a2 7z- , ( 2 7 2 ) %a + n(2.35) where P is the wheel load in kN and a is the radius of the loaded area in meters and(2.36)where p is the average contact pressure which is assumed to be equal to the tireinflation pressure in kPa for a single wheel or 0.7 to 0.8 times the air pressure of dualtires.The permissible subgrade stress, which is expressed as a function of theundrained shear strength of the subgrade, is calculated using the following equation :o- p, = A 2,^ (2.37)where A is a dimensionless coefficient varying between 1.9 and 2.0 governed by thelateral restraint provided by the geotextile.Chapter Two - Literature Review^ 37Vertical support from the geotextile is described by a differential stress across thefabric, assuming a circular deformed shape of the geotextile-subgrade interface shownin Fig. 2.4.of wheal pathFigure 2.4 Assumed deformed shape of base course-subgrade interface(Kinney and Barenberg, 1982)The differential normal stress in kPa across the geotextile istoAo-z =Rwhere tg is the tension in the geotextiletg = Eg Eg (kPa)(2.38)(2.39)and eg and Eg are the geotextile strain (%) and geotextile modulus (IN/m/% strain),respectively.The average strain in the geotextile is determined from the assumed deformedcircular shape according to the following equation:4,rRO=^ 2g 135B'(2.40)Chapter Two - Literature Review^ 38where^ R=^sin 9and^ 0= 2 arctan(-5d3H(2.41)(2.42)Design charts can then be produced based on the above procedure, where the appliedstress on the subgrade surface is set equal to the permissible stress plus the differentialnormal stress in the geotextileoz = Cr per + AO z_g (2.43)and the expression solved for various values of the subgrade undrained shear strength.Giroud and Noiray (1981) developed a design procedure based on theoreticalconsiderations and data from field trials. Their approach has become the basis formany design methods, and is described in detail below. It includes a tensionedmembrane effect and incorporates, for the reinforced system, a bearing capacity failurein the subgrade that is one of general shear failure, rather than a local shear failure.Essentially, failure is anticipated if the sum of the surcharge pressure on thesubgrade below the loaded area, minus the pressure reduction due to the tensionedmembrane effect, exceeds the ultimate bearing capacity of the subgrade, that isqu = (x+ 2)c. + yh= p– pg (2.44)where qu is the ultimate bearing capacity of the subgrade, p is the surcharge pressureon the subgrade and pg is the reduction of pressure due to the tensioned-membraneeffect.The failure of the unpaved road can occur in the subgrade or the geotextile. The basecourse material is assumed to have sufficient friction to ensure mechanical stability ofthe layer and to prevent sliding over the geotextile. A pyramidal load distribution isChapter Two - Literature Review^ 39assumed (Fig. 2.5). For a base course thickness of h o without a geotextile, thepressure at the base of the base course layer is given by:2LBpecPo = 02+2ho tana0 XL + 2ho tana0 ) ± 7h° (2.45)A base course thickness of h in the reinforced case, yields a stress on the subgrade of:2 LBP ec P =^ F2(B + 2h tan ce)(L + 2h tan a)(2.46)(a) (b) ASubgrade2a280^B 0GeotextileFigure 2.5 Load distribution by base course layer: (a) Case without geotextile; and(b) Case with geotextile (Giroud and Noiray, 1981)The axle load, P, is evenly distributed between the two wheel sets and assuming dualwheels:P = 4 AcPc (2.47)where A, is the contact area (m2) and pc is the tire inflation pressure (kPa).In the method each dual wheel contact area 2Ac is replaced by a rectangle of alarger area of B times L according to following relationship:LB = 2 AcAri^ (2.48)Chapter Two - Literature Review^ 40Two different cases are examined, on-highway trucks and off-highway trucks. For atire inflation pressure of p c the following approximate relationship between L and B isintroduced:B^B =L (on-highway trucks)^(2.49)V2 PcandL B = ^ (off-highway trucks)^(2.50)PcThe maximum subgrade pressure in the unreinforced case is equal to the elastic bearingcapacity, which is the same as local shear failure value according to Terzaghi:Po = cu r. ho (2.51)Using the above relationship and equations 2.49, 2.50 and 2.45, the base courseaggregate thickness h0 in the unreinforced case can be determined from the followingequations:PCu = 22TP— +2ho tan ao ^+2ho tanao01Pc^2pc(on-highway trucks) (2.52)Cu =P (off-highway trucks) (2.53)27r ^ +2ho tan ao411 P an212,^+ 2ho t aoPcA^ls_.. ,\\(P) --)- --Initial locationr13Th/Z:I!^/AGeotextileof geotextile et^es2a 2a' 4^ 2avChapter Two - Literature Review^ 41The subgrade soil is assumed to be fully saturated, and therefore incompressiblein the undrained state under traffic loading.Figure 2.6 Assumed parabolic shape of deformed geotextile(Giroud and Noiray, 1981)Strain in the geotextile is calculated, for an assumed parabolic deformed shape, fromthe following:b^ '6= +b 1^for a' > a^(2.54)a+ a's= b— -1afor a > a'^(2.55)where b and b' are the half length of (P) and (P'), respectively, and the widths a and a'are obtained from Figs. 2.5 and 2.6:2a=B+2htana^(2.56)2a' = e — B — 2h tan a^(2.57)Chapter Two - Literature Review^ 42The chord lengths b and b' are calculated using the following relationship between thearc of a parabola and a chord:ba1 = 12-2± a in {2s +111412s)}_ 2 12s^a^)2+ (2/(a\\ 2__ 1.b'^1[1i + (2(r –s)) 2^a' ^{2(r – ^1112(r s))a'^2^a'^2(r – s)^a'1ln + 1+^_a(2.58)(2.59)The reduction in pressure due to the tensioned-membrane effect, p g, isconsidered to be a uniformly distributed pressure on AB (Fig. 2.6) and is equal to thevertical component of the tension T in the geotextile at points A and B:apg = Tcos/3^(2.60)Knowing the strain e , and the geotextile elastic modulus E, the tension is given by:T=Es eand from the following property of a parabola:tanf3=—a2s(2.61)(2.62)then pg is established from:E ePg 2aill+(±1–)2s(2.63)The thickness of the base course layer in the reinforced case is then determined fromequations 2.44, 2.46, and 2.63:P E e(g+2)cu . ,^ (2.64)20' +2h tan a)(L + 2h tan a)^± (a / ) 272sChapter Two - Literature Review^ 43The reduction in the base course aggregate thickness due to the action of the geotextileis then given by:Ah=ho — h^ (2.65)using equation 2.52 or 2.53 to establish hp, and solving equation 2.64 for h. The valueof reduction thickness, ih, is assumed to be independent of the traffic volume.In order to account for traffic loading Giroud and Noiray (1981) proposed aformula based on work by Webster and Alford (1978) to determine the required basecourse thickness of an unreinforced road:0.19 log N no CBR 0.63(2.66)where Ns is the number of passes of a standard axle load, P s = 80 kN. This equation isrestricted to a standard axle load and to a rut depth of 0.075 m. It was suggested otheraxle loads be accommodated through the widely used power rule in unpaved roaddesign:N _(P;Ps)3.95(2.67)where Ni is the number of passes of axle load Pi. Equation 2.67 was extended to rutdepths other than 0.075 m using an expression, deduced from the Webster andWatkins (1977) test results, where log N s was replaced by:log A r — 2.34(r-0.075)^(2.68)The unreinforced base course thickness can now be determined for traffic loading bycombining equations 2.66, 2.67 and 2.68, and substitution ofcu (kPa) = 30.(CBR)^(2.69)Chapter Two - Literature Review^ 44giving:1.61935.1ogN+ 6.39642- log P — 3.78927.r — 11.88877no =^ 0.63^ (2.70)CuThis equation is not recommended for more than 10,000 vehicle passes.Finally, the thickness of the base course in the case of a reinforced system, againconsidering the influence of trafficking, given by:= 0 — Ah^ (2.71)Use of these equations was simplified by creation of design charts. It should benoted that the rut depth, r, is defined as the vertical distance between the highest pointof the road surface between the wheels and lowest point of the rut (Fig. 2.6).Giroud et al. (1984), modified the above design approach for use of geogridsrather than geotextiles to take into account interlocking of the geogrid and base coursematerial. Progressive deterioration of the subgrade shear strength is described by:= cuAr =^1 cu^[1 ± (log N) 15 Cu 11000(2.72)Progressive deterioration of the base course layer is taken into account by decreasingthe load distribution angle with increasing number of passes. The load distributionangle cco is determined by combining equations 2.46 and 2.49, 2.51, 2.66 and 2.72,resulting in the following equation:2p,^ ,krcu^\\ Ps 2p110.17157 ^ + 2Ps (2.414211.1tan ao = 6.5•loog3 NC .6 u(2.73)Chapter Two - Literature Review^ 45The design is the performed according to the same procedure described by Giroud andNoiray (1981), with a recommendation that the tensioned-membrane effect beneglected or if the membrane effect is not negligible a lump reduction of 10% of thedesigned base course layer is recommended.Sellmeijer et al. (1982) developed a design method based on the tensionedmembrane theory. Different from other available design procedures, their methodfulfills equilibrium conditions for both the membrane action of the geotextile and thesubgrade soil. Therefore the deformed shape of the geotextile-subgrade interface is notdefined arbitrarily which is contrast to other semi-empirical methods.When a geotextile is subjected to a traffic load q c, the geotextile starts to deformand goes into tension, hence developing the tensioned membrane action. Theequilibrium equations are:T = Th kds )+12I I(2.74)desTh dx2^ql)where T is the tensile force in the geotextile, Th is the horizontal component, s is thevertical displacement of the fabric and x is the horizontal distance from the truck axis.The subgrade soil is assumed to behave in a bi-linear elasto-plastic manner, where thelinear elastic phase is described by using a subgrade reaction coefficient, C c :q = Cc- s (2.75)In the rigid plastic phase the subgrade behavior is described based on Brinch Hansenbearing capacity formula:q1 = N,(c+ 4 Wy. tan2^(2.76)where Nc is assumed to be 5.14 and di is the internal friction angle of the subgrade.Chapter Two - Literature Review^ 46This method is two dimensional and the distributed traffic load is assumed to beconstant acting over the whole length of the road structure given by the followingequation:Pq0 = i20B+ 2eh)(B + 2eh)(2.77)nBnB + 2eh v 1^ h ___ Geotextile 14^ W/2 Figure 2.7 Road cross section showing pertinent factorsSolving equation 2.74 for the constant load distributed by the base course layer givesthe following:Cc s 2 cosh(fl— a) cosh sinh 77=qo^sinh /3Cs _ 2 cosh(fl— a) cosh sinh 77470 — sinh fi + 1— cosh(— a+ 77)Cs . 2 cosh a cosh( —A sinh 77go sinh 13(2.78)(2.79)(2.80)A/2iChapter Two - Literature Review^ 47where:a= 1 A.11C,^Cfi = _} w . li cThTh77= 2(nB +2eh). C1 -^= x - II-C-LTh ThEquation 2.78 is valid for 0 < k < a-ii, equation 2.79 is valid for ot-ri < k < a-Frl, andequation 2.80 for a-Fri < k < 13.The width of the plastic zone, b, is obtained from equilibrium considerations ofthe subgrade reaction to the sum of the wheel load and the base course layer aggregateweight:Pbql= 2(B +2eh) + 241bOr^ (2.81)b = P2(q1 — rh)(B +2eh)The three possible locations of the plastic area of the soil stress are shown in Fig. 2.8,in which the occurrence of the first case is the most frequent one.Equation 2.74 can now be solved together with equations 2.76 and 2.81 for eachof the three cases. The horizontal tension force, Th, is however still unknown but canbe determined by the compatibility condition of the geotextile if its anchorage isconsidered being sufficient.V V V' ♦ Y♦ ®Chapter Two - Literature Review^ 48i•V\"-TYVbW/2k^ b )1 4 N2♦ V♦MIIMMINIIIIIM14^ b ofFigure 2.8 Location of the plastic zonesThe ultimate solution of the membrane equation is given in the following way:(b — nB — 2eh ± d) 2^As= ^4 (b—nB —2eh)21^2E g^(b — nB —2eh) 2 + 3d2 32= ^[( 3(q1 — 741)W)^b27;i= +(qi yh)b (b —nB b—^222eh+d) +2.2(2.82)(2.83)(2.84)In the above equations, As is the total rut depth, Eg is the geotextile Young's modulusand Tm is the maximum tensile stress in the geotextile.Chapter Two - Literature Review^ 49The location of the plastic zone is defined by the plastic location parameter, d:d = 0^if b < A and b < W - A (case a)d = W - A^ifb>AandbW-A (casec)This design method does not take traffic into account, but based on theMoordrect test results, described by Sellmeijer and Kenter (1983), the followingequation is suggested to calculate an equivalent axle load Pi for N passes of a real axleload Ps := p6.2 (2.85)Haliburton and Barron (1983) presented a design method based on a scale modellaboratory tests and previous research by Haliburton et al. (1980) and Haliburton andLawmaster (1981). The approach focuses on placement of the reinforcement at anoptimum depth within the base course layer, given the width of the loaded area.Research showed that maximum restraint of the base course material occurred whenthe base course aggregate thickness was equal to 0.33B, where B is the width of theloaded area. This optimum depth was found to significantly increase the ultimatestrength and deformation resistance of the base course material as well as showing asignificant stress reduction at the subgrade interface of 50%, based on a Burmisterlayered system theory (Burmister 1943, 1944).In design, knowing the axle load and the tire inflation pressure, the contact areais established, either as circular or rectangular. Then, by applying Boussinesq theorystresses are calculated at a depth of 0.5B tan 4), or alternatively at a depth of 0.33B,which are considered comparable. The predicted stress is then reduced by half toreflect stress reduction effects observed in the research. The reduced stress isChapter Two - Literature Review^ 50compared to the ultimate bearing capacity of the subgrade, and if it smaller the designis satisfactory.Milligan et al. (1989) presented a design method for designing unreinforced andreinforced unpaved roads based on extensive tests on small and large scale modelsunder monotonic and cyclic loading, and a large strain finite element analysis. Theirmethod does not follow the conventional tensioned-membrane approach, butemphasizes the role of shear stresses at the subgrade surface. When load is applied tothe base course layer it produces vertical and horizontal stresses in the subgrade layer.Some of the horizontal stress is resisted by the base course material outside the loadedarea, and the remainder develops outward acting shear stresses on the subgradesurface. This action reduces the subgrade bearing capacity significantly. If ageosynthetic is installed at the subgrade surface, the outward acting shear stresses willbe taken up by tension in the geosynthetic, promoting a vertical stress only to betransmitted to the subgrade soil.The method, which is developed for use in design of unpaved roads at small rutdepths, does not consider a tensioned-membrane effect and anchorage is considered tobe less important than in other procedures. However, stiffness of the geosynthetic isimportant if small ruts are desirable. The authors recognize the importance of thetensioned-membrane effect, but only at considerable rut depth. The bearing capacity ofa strip footing is predicted by using upper and lower bound theorems of plasticitytheory (Fig. 2.9):2Nca =l+--7r +arccos(-L)+ — (—L-1^(2.86)2^s^swhere —Nca = \"SuChapter Two - Literature Review^ 51Cr Vati a Hill/ \\BFigure 2.9 Bearing capacity of subgradeThe load spread angle [3 is used to estimate the vertical stress within the basecourse layer. It should be noted that B is taken as the half-width of the footing width(Fig. 2.10). The vertical stress p is applied to the footing of width 2B and resultingstress at depth z below the base course surface is given by:pB=^a +B + z tan f3 (2.87)14^ B 64 B'Figure 2.10 Load spread angle below plain strain footingavoUnderneath the footing the base course is pushed outwards, developing activepressures, and passive pressures are developed outside the influence zone of loading.hKafav dz KpY D2A^B\\^CD^ EChapter Two - Literature Review^ 52Taking into account the angle of friction 5 on the base of the footing, the minimumaverage shear stress required on the subgrade surface for stability is determined bysolving the horizontal equilibrium of the rectangular area shown in Fig. 2.11:Tr=-1(Ka-Kp)()/.. tanflB' )472 Kap ( B log( B)._ p(--B) tan 8 (2.88)This equation gives the linear increase of the required shear stress with vertical loadshown as the \"required\" line intercepting the available shear stress at point A inFig. 2.12.pB tan 8s r B'Figure 2.11 Soil block in equilibrium analysis27C+2 TsuFigure 2.12 Normal and shear stress interaction at subgrade-base course interfaceChapter Two - Literature Review^ 53Limiting equilibrium is achieved when the required subgrade bearing capacity matchesthe available subgrade bearing capacity. The applied vertical stress can be calculated,once the mobilized bearing capacity is determined, using the following equation:p=Arcc.(1+ — tanfi) (2.89)Bearing capacity failure can also occur in the base course layer, and it is thereforenecessary to check that the axle load is not limited by the bearing capacity of the basecourse.Design using this method requires knowledge of the half-width of the loadedarea, the base course thickness, value for the load spread angle, base course frictionangle and unit weight, and the undrained shear strength of the subgrade. Theinterception point, A, is determined graphically or numerically. The applied verticalstress, p, is then calculated and the required reinforcement force in the geosyntheticdetermined knowing that T = TrBt.Interpretation of the results established that the greatest benefit of areinforcement is obtained when using a poor base course material on very softsubgrades, a normal strength base course on relatively firm subgrades, and strong basecourse on strong subgrades.Houlsby and Jewell (1990) extended the procedure in to cases of axisymmetricloading, though the modification complicates the calculations to an extent that may beunnecessary given the nature of the application.Chapter 3Test Site Description3.1 GEOLOGICAL HISTORY3.1.1 Geological History of the Lower MainlandThe test site is located on Lulu Island in the Fraser Delta of the Lower Mainlandof British Columbia. The geological history of the Lower Mainland has been studiedby Armstrong (1956, 1957, 1984), Blunden (1973, 1975) and Clague and Luternauer(1983), who reported that the bedrock geology of this area has evolved over millionsof years and the unconsolidated Quaternary sediments over the last several thousandyears. Since the bedrock at or within 10 m of the surface forms less than five percentof the Fraser Lowland, the main concern for shallow foundations is the Quaternarysediments.The Fraser Lowland experienced at least three major advances and retreats ofglaciers during the Ice Age, and the present surficial geology of the Vancouver area is54Chapter Three - Test Site Description^ 55strongly influenced by these repeated Pleistocene glaciations. Each major glaciation inthe lowland, which was accompanied by isostatic and eustatic adjustments, wentthrough three main stages (Armstrong, 1984); an advance stage, a maximum stagewhen ice attained a thickness of 1800 m or more, and a retreat or deglaciation stage.As a consequence of the complex deposition environment, the sediment types rangefrom glacial to deltaic and demonstrate considerable heterogeneity both laterally andvertically.The Fraser Lowland, also known as the Whatcom Basin, is a triangular area ofapproximately 3500 km2 that forms the southwest corner of the mainland BritishColumbia and the adjoining northwest corner of Washington State (Fig. 3.1). TheLowland is bounded on the west by the Straight of Georgia, on the north by the CoastMountains and on the south-east by the Cascade and Chuckanut mountains. For thelast 100 millions years the geology of the greater Vancouver area has been that of theinfilling of this structural basin which originated during the Upper Cretaceous time.Figure 3.1 Index map showing location of Fraser Lowland.Chapter Three - Test Site Description^ 56Approximately 11,000 years ago, after the ice sheets of the Fraser Glaciation hadretreated from most of the low lying areas of southwestern British Columbia thesurface of the Lower Mainland, which had been depressed by the weight of the ice,rose 150 meters (Eisbacher, 1977). Following the complete withdrawal of the ice andthe adjustment of the level of the land some 10,000 to 8,000 years ago the FraserRiver, which is the most important agent for the deposition in the region begandischaring into the Strait of Georgia. Currently it discharges approximately 18 milliontonnes of sand, silt and clay each year (Armstrong, 1984), most of which is depositedin the Strait of Georgia to add to the growing delta.Due to the large volume of sediments supplied to the river by the retreating icesheet, the delta expanded very rapidly and attained its present position about 5,000years ago. The advance of the delta continues to this day, causing the mouth of theriver to move seaward at an average rate of 2 meters per year near the water surface.At a depth of 100 meters the advance of the delta front is as much as 9 meters per year.3.1.2 Surficial Geology of Lulu IslandThere are a wide variety of Quaternary deposits found in the Fraser Lowland.The three major sediment groups are: waterlain sediments; glacier ice sediments; andglaciomarine and glaciolacustrine sediments. In addition to these sediments occasionaldeposits of windblown material are found. Peat bogs and organic sediments are verycommon in flat lowland areas.Lulu Island and adjacent Sea Island, see Fig. 3.2, are only a few thousand yearsold and both owe their existence to the growth of the Fraser Delta complex. Theylikely originated as sand bars at what was then the mouth of the Fraser River, some8,000 years ago. Approximately 1000 years later the area had become a mud flat andsalt marsh, and the Fraser River had developed what are now termed the North andChapter Three - Test Site Description^ 57South Arms. Lulu Island grew to its present size 4,000 years ago, at which timeextensive peat bogs had formed upon the earlier salt marshes. These peat bogs nowcover about 14 km2 . The top layer of the peat bog is referred to as unhumified peat,which is a dead sphagnum moss that is only slightly decompressed.The current surficial deposits of Lulu Island comprises a thin discontinuouslayers of clays, silts and peats, which have been laid down in fluviatile, interdistributaryswamps, and salt marsh environments. They overlie a thick sequence of sands andsilts. The bedrock is of Tertiary age and is found at depths of between 200 and 270meters.3.2 DESCRIPTION AND LOCATION OF TEST SITEThe research site is located on the eastern part of Lulu Island, see Fig. 3.2. It ison ground which is owned by the Province of British Columbia Ministry ofTransportation and Highways, being on a right-of-way parallel to and southeast ofHighway 91/91A, some 30 km from UBC.The site is bounded to the north and west by Highway 91/91A, and to the southand southwest by a main drainage ditch that runs due north, see Fig. 3.3. The southside, between the site and Hamilton Road, is an agricultural land. The site area isapproximately 100 m long and 10 m wide, and located about 350 m east of theHamilton Interchange. It forms part of a future highway development area, but isundeveloped, being covered by grass overlying a 30 to 60 cm thick sand layer that wasthe edge of the preload for existing highway.Chapter Three - Test Site Description^ 58Figure 3.2 Location of the research site.Figure 3.3 General research site details.The ground surface elevation at the site is about + 1.0 m with respect to theaverage sea water table. The thickness of the surficial soils, which generally compriselayers of peat, organic silts, and clays, silts, and sandy silts, is about 15 m; these soilsare underlain by a fairly dense sand or silty-sand . The depth of peat formation in thisChapter Three - Test Site Description^ 59part of Lulu Island varies greatly from 0 m to 6.1 m (Kern and Buchanan, 1984). Theelevation of the groundwater table as determined from the ditchwater levels variesbetween + 0.0 to + 0.5 m.3.3 SUBGRADE SOIL PROPERTIESThe final location of the test site was selected to give a reasonably uniform nearsurface stratigraphy along the unpaved road test section, which was 80 m long and 6 mwide. Following a review of borehole logs in the general area, a detailed siteinvestigation was conducted, that included test pits, laboratory tests and field vanetests. The site investigation was done to establish the uniformity of soil stratigraphyand to determine the variation in physical properties of the subgrade soil both spatiallyand with depth. Locations of the test pits and sampling points at the final location areillustrated in Fig. 3.4. A total of 9 test pits were dug that varied in depth from 0.5 m to2.0 m, and 10 field vane tests were performed. Details of the site stratigraphy, soilproperties and variation of shear strength with depth are given below.3.3.1 Site StratigraphyThe description of site stratigraphy is based on examination of the nine test pits,six of which were relatively shallow and used to determine the depth to the softsubgrade and to locate the groundwater table. The other three are deeper and providemore detailed information.The characteristic sequence of soil layers is illustrated in Fig. 3.5. The sequenceis an average of the three deep test pits and the elevation shown is based on the localelevation system in the City of Richmond. The vegetated topsoil overlies a slightlyField Vane TestsUndisturbed Samplesfor Triaxial TestsElevation0.80 m —0.75 m — Topsoil, turfLoose, moist, grey brown, fine to mediumslightly silty SAND, traces of organics0.50m^ (SWM;SW-SM )Soft to firm, grey mottled reddish- or yellowish-brown, sandy, very clayey SILT of high plasticity,traces of otganics(MIICS;MS-MC)Grade 0.06 mlevel 0.10 mSoft, dark brown mottled grey, very organic,very clayey SILT, of extremely highplasticity(MECO,OH)(British ; Unified) - Soil Classifiacation SystemGroundwater table^•K?..;:;:7- 0.20 m —^Soft to firm, grey mottled brown, organic,slightly sandy, very clayey SILT, of veryhigh plasticity(MVCSO;OH)Chapter Three - Test Site Description^ 60silty sand and a yellowish-brown sandy, very clayey silt, both of which showed somevariation in thickness across the site.Figure 3.4 Location of testsFigure 3.5 Typical test pit resultsChapter Three - Test Site Description^ 61Below this was an organic very clayey silt, the upper part of which was slightly sandy.Uniform thickness of the upper soil layers, which varied in thickness from 0.0 up toabout 0.6 m and was interbedded in some places with 5 - 20 cm thick sand lenses.Prior to construction of the test sections, these upper soil layers were excavatedto the grade level (see Fig 3.5) to promote a uniformity of soil stratigraphy along thelength of the trial.3.3.2 Soil PropertiesProperties of the subgrade soil were established from grain size analysis andAtterberg Limit tests. Samples were taken at 0.2 m intervals below grade from thethree deeper test pits, the locations of which are shown relative to the road trial inFig. 3.4.Specific gravity values were obtained from two different depths from each testpit, at 0.1 m and at 0.9 m depth. Since soil in the field composed of only one type ofsoil or clay mineral is rarely found, but more often the soil is composed of variouspercentages of sand, silt and clay, the specific gravity value of the solids obtained bythe specific gravity test procedure is actually an average of all the types of mineralsoccurring in the soil being tested. Results from the specific gravity test performed areshown in Table 3.1.Table 3.1 Specific Gravity ResultsTP1GsTP2GsTP3GsAverageValue2.656 2.573 2.690 2.640100908020100Chapter Three - Test Site Description^ 62The grain size distribution for the subgrade soil was determined from the abovementioned test pits at 0.2 m interval using the average specific gravity values fromTable 3.1. A hydrometer method, which depends on Stokes' equation for the terminalvelocity of falling sphere, was used for this purpose. A total of 24 hydrometer analysiswere performed to establish the grain size distributions, see Fig 3.6. The shaded area isthe result from all the test performed for the two subgrade soil layers, where the lowerbound of the shaded area represents the slightly sandy material, MVCSO, and theupper bound represents the MECO material which starts at depth of 0.26 m, seeFig. 3.7 for more details.SandFineSilt ClayCoarse Medium Fine Coarse Medium Fine0102030^ co4050^C6070a_80901 0001 0.01 0.001 0.0001Grain size, in mmFigure 3.6 Particle size distribution curve - Hydrometer analysisA knowledge of the range of moisture content over which soil will exhibit acertain consistency is often beneficial because the behavior of fine grained soils is oftenChapter Three - Test Site Description^ 63related to the amount of water contained in the pore spaces. After a cohesive soil hasbeen remolded, its consistency can be changed by increasing or decreasing the watercontent, and the water contents that correspond to the boundaries between the statesof consistency are called the Atterberg limits.The variation of plastic and liquid limit, and the natural water content of the soil,with depth below grade are reported in Fig. 3.7. The values were determined at 0.2 minterval, and represent average values from the three deeper test pits. The values of allthree indices increases uniformly to a nearly constant value at a depth of 1 m.Soil Type^Water content (%)Grade level0.06 m^0^100^200^300 01111Iii1111111111F wp (avrg.) - awp_^(avrg.) + awl0 Plastic Limit (average), wp• Natural Water Content (average), wn^ Liquid Limit (average), w10.0 0.2 —0.4 —0.6 —0.8 —1.0 —1.2 —1.4 —1.6 —1.8 —2.0 (%)Figure 3.7 Atterberg limits and indices - Test Results'^I^'1 11111 11 1^1111 1 111^I^I• TP1• 1P2• TP3100^150^200^250^300Liquid Limit (w 1 )0• •• ••1505041.010*P.7kkilgth higniessFeE+e6Elib\" !^ Vekrxis Clam p Peavorei:CCMIAP.ARINCliSOltfiDUAL Loma LIMIT- • — 1 , 1 ,,^\\y,wow* Rabe ofChapter Three - Test Site Description^ 64The results indicate the subgrade material is highly plastic (LL > 50%). Thewater content is very close to the liquid limit throughout the profile, but does notexceed it. A water content in the lower subgrade layer greater than 100 % suggeststhat the subgrade is very organic.Figure 3.8 Relationship between liquid limit and plasticity index for the subgrade soilThe same data are illustrated on a plasticity chart in Fig. 3.8. The plasticity chartis based on the British Soil Classification System (BS 5930:1981) and is devided intofive ranges of liquid limit where the four highest ranges (I, H, V and E) can becombined as the upper plasticity range or the high plasticity range for the Unified SoilClassification system. The A-line on the plasticity chart is used to differentiate betweenclay (C) and silt (M), where silt plots below the A-line and clay above, which meansthat silts exhibit plastic properties over a lower range of water content than clayshaving the same liquid limit. If fine grained soils contain a significant amount ofChapter Three - Test Site Description^ 65organic material they usually have high to extremely high liquid limits and plot belowthe A-line as organic silts. A peat typically exhibits water content greater than 350%.The liquid limits for most of the samples taken from the subgrade soil have extremelyhigh liquid limits. Those with a liquid limit between 50% and 90% are samples takenfrom the upper subgrade layer shown in Fig. 3.7, between 0 and 0.26 m depth, andthose with a liquid limit close to 300% were taken at a depth of 1.1 m. Thisinterpretation of the tests indicates the subgrade soil to be an organic clayey silt, whichis in a good agreement with the grain size analyses.3.3.3 Shear Strength CharacteristicsThe undrained shear strength of the subgrade material was determined using afield vane: locations of the seven profiles within the test section area are shown inFig. 3.4. For comparison purposes, unconsolidated undrained triaxial tests were alsoperformed, and the locations of the undisturbed sampling for these UU-tests are alsoshown in Fig. 3.4.The field vane is a widely used in-situ test for evaluating the undrained shearstrength of fine-grained soils. Originally used in Sweden in late 1920s, it has beenemployed extensively worldwide since the late 1940s. Even though the field vane hasbeen used for over 70 years there are still numerous questions that still remainunanswered about the interpretation of the test. Among the various factors that affectthe interpretation of the test, it is recognized that shear stress distribution around thevane, rest period following vane insertion, fabric disturbance, pore-pressure buildup,strain rate effects, progressive failure and shear strength anisotropy have majorinfluence on the measured strength (e.g.: Chandler, 1988; Roy and Leblanc, 1988;Becker et al., 1988; Silvestri and Aubertin, 1988).Chapter Three - Test Site Description^ 66The field vane tests were performed at 0.5 m interval for the first 4.2 m depthand thereafter at 1 m intervals. The equipment used was a Swedish Vane Borer likethat developed at Chalmers University of Technology in Sweden, manufactured byNilcon. The main advantage of the Swedish vane is that it needs no protective casingaround the vane and the rods, which makes the operation much faster. Also, in softground, no pre-bored hole is required. A disadvantage of the vane is that there is noguarantee that the rods will advance into the ground without bending. The vanes usedwere 13 cm high and 6.5 cm in diameter and 17.2 cm high and 8 cm in diameter.The vane borer consists of an instrument for applying torque, supported by aportable frame. The 20 mm diameter rods pass through the torque recorder and areconnected through a slip-coupling to a vane. The vane is advanced into the ground bya system of hand cranking and as the vane advances into the ground the slip couplingallows the vane to rotate in a clockwise direction through approximately an angle of15° in order to measure the soil-rod friction. A torque is then is applied at the rate of0.2°/s at the top of the rod (2 handle rotations/s in low gear), and recorded withrespect to angular rotation on a sheet of waxed paper in the torque recorder. Theresults are used to calculate the shear strength of the soil, knowing the differencebetween torque required to turn the rods plus vane, and torque required to turn therods alone. All field vane undrained strengths were calculated using the conventionalexpression (Greig et al., 1988):6McU = 77rD3 (3.1)wherecu = undrained shear strength,M = applied torque, andD = diameter of the vane.600^10 20 30 40 50iiit1111111111111,,11111111,,0 2 4 6 8 10Ifiliiil,11111,111, [1111[1111111111111c„ (kPa)0 4 8 12 16c uu (1cPa)Chapter Three - Test Site Description^ 67It has been noted that the waiting period following vane insertion has a verysignificant effect on the shear strength and in order to minimize the influence of thistime lag between intrusion of the vane and shearing a delay time of less than 1 min wasretained throughout all the tests.Undrained Shear Strength^Residual Shear Strength SensitivityFigure 3.9 Field vane shear test resultsResults of the vane tests are shown in Fig. 3.9. The solid line on the figuresrepresents the average value from all of the tests. The profile of strength with depth isnearly uniform, and between 30 and 40 kPa. There is a slight increase in strength nearChapter Three - Test Site Description^ 68grade level which is attributed to the lower water content of the soil in this zone.Apart from this zone, below 0.5 m the consistency of the subgrade according to Britishclassification is soft, that is a undrained shear strength between 20-40 kPa.The sensitivity of the subgrade soil, which is defined as the ratio of the undrainedstrength in the undisturbed state to the undrained strength at the same water content inthe remolded state, is fairly uniform through the first 4 m in depth. The averagesensitivity is about 6. Clays with sensitivity between 4 and 8 are described as sensitive,and those with sensitivities between 8 and 16 are described as extra sensitive.Field vane corrections and correction factors have been developed throughcomparison of field vane strength and the strength back calculated from actual failures,and from laboratory tests. However, most engineers use eqn. 3.1 when calculating theshear strength, even though it results in some conservatism. No correction factorswere applied to the vane data reported in Fig. 3.9, and it is felt the data may slightlyunderestimate the shear strength.Results from the unconsolidated undrained triaxial tests were in good agreementwith the field vane tests. Undisturbed samples for the UU-tests were taken at about10 - 25 cm below grade level at the locations shown in Fig. 3.4. A total of nine UU-tests were performed, three from each hole, giving an average undrained shear strengthvalue of 41 kPa.Chapter 4Description of Test Sections4.1 TEST SECTION LAYOUTThe test site comprises five test sections. One is unreinforced, three sectionsinclude a geotextile and one includes a geogrid. Each was trafficked in sequence by avehicle of known axle loading. Design methods proposed by Steward, Williamson andMohney (1977), Bender and Barenberg (1978), Giroud and Noiray (1981), andHaliburton and Barron (1983), were used to estimate the required thickness of the basecourse layer. The basic design criterion, apart from the subgrade properties, was thenumber of passes to be applied by the vehicle: the sections were intended to have alifetime of a minimum of 100 passes and a maximum lifetime less than 1000 passes.Lifetime is defined with respect to a serviceability type of failure, and taken to bea rut depth at which the vehicle can no longer traffic the structure. An importantgoverning parameter for each geosynthetic is the influence of base course thickness on69Chapter Four - Description of Test Sections^ 70the relationship between number of passes and rut depth. Therefore each sectionincorporated a variable thickness of base course aggregate.The arrangement, construction and instrumentation of the test sections aredescribed in detail below. Material properties of the base course aggregate andgeosynthetics are reported in Sections 4.2 and 4.3, respectively.4.1.1 Arrangement of the Test SectionsThe five test sections were arranged in sequence, see Fig. 4.1, to facilitate trafficloading. The arrangement was selected to accommodate constraints of the site andpromote ease of construction. Each test section is 4.5 m wide and 16 m long. Thefirst section is the unreinforced section and serves as a control section in the study.The next section is reinforced with a geotextile, Texel Geo 9 and the section thereafterwhich also includes a geotextile, is the Polyfelt TS 700. The fourth section includes ageogrid, Tensar BX 1100, and the last test section is has a geotextile, Polyfelt TS 600.Details of the base course layer are shown in Fig. 4.2. The thickness varies ineach test section between a minimum value of 25 cm and a maximum value of 50 cm.The 25 cm layer is 3 m long, and increases in thickness to 50 cm over a length of 5 m,giving a slope gradient of 5 %. This longitudinal geometry was based on the area atdisposal, specifically the total length, the length of the loading vehicle, and a desire toassess the repeatability with regard to the layer thickness within each section.Therefore, instead of only taking measurements during traffic loading at one basecourse thickness in each test section, it allowed two measurements, which greatlyimproves the quality of the data by allowing the repeatability and uniformity of anindividual section to be quantified.Chapter Four - Description of Test Sections^ 71—A 480 m—4—- 4.5 m -Am— - Umeinforced Tema Geo 9 Polyfelt TS 700 Tamar BX 1100 Polyfelt TS 600 -16 m^16 m^16 m^16 m^16 m,1(Plan view of test roadNOT TO SCALELongitudinal cross sectionA - Awe figure 4.2far detailsFigure 4.1 Test section layoutm^5 moak m^► 1-4---3 m—e...1 •^55% 25 cm^ 5%Base courseGeosyndtetieSubgradeH Longitudinal cross sectionScale 1 : 125Figure 4.2 Test section geometry4.1.2 Instrumentation and MeasurementsPerformance of the test sections was evaluated from measurements of:• rut depth;• base course thickness;• base course deformations;• geosynthetic strain; and• deformed profile of the geosynthetic,with increasing number of vehicle passes.1\\50 cm16 mChapter Four - Description of Test Sections^ 72Rut depth measurements were taken across ten different stations, see Fig. 4.3,defined by the two locations in each test section where the base course thickness was50, 40, 35, 30 and 25 cm.Is-1.5 molls 5 m^ 3 m ^►14^5 m^m0.5 m^4.0 m 3.5 msee figure 4.4 for detailsLocation of settlement plates[I^ 16 m ^a^0.5^9.0^1.0^1.0^1.5^2.0^1.5^1.0^1.0^3.0^0.5I r1 ■•1 Poi^WI^PO^01 ■••1Location of marker suds for strain measurements = •and locations of rut depth measurements =Scale 1:1251■11C, ■1111111111•Ic/c 30 cm c/c 15 cm c/c 30 cm c/c 15 an c/c 30 cmt t^tt tttttt t^t^tttt t tttt^t^t4.5 m ^ ►Studs location transverseon road alignmentScale 1:50Figure 4.3 Instrumentation detailsA sophisticated survey network was established to expedite the recording ofthese values, together with the development of surface rutting with cumulative vehiclepasses. At every measurement station, wooden pegs were driven into the ground onChapter Four - Description of Test Sections^ 73both sides of the test road. The tops of the each pair of pegs at every station wereplaced at the same elevation, using a nearby bench mark and the local elevation system.To measure the rut depth at every station a string was stretched between the pegs and,knowing the elevation of the pegs, a measuring tape was used to measure the distancefrom the datum elevation to the base course surface. The same datum was used toestablish the deformed profile of the geosynthetics after trenching down to exposethem at selected locations. The elevations of the pegs were checked periodically toensure they were not disturbed by the vehicle passes. Figure 4.3 shows the locations ofthe rut measurements, which is at the same locations as the strain measurements,except in the middle where the rut measurement is above the settlement plate in the 25cm thickness section.The thickness of the base course was established from measurements usingsettlement plates incorporating a special tube that passes up through the base courselayer. They were installed at three different locations within each test section, at basecourse thicknesses of 50, 35, and 25 cm to measure both the settlement of the subgradesurface and the thickness change directly in the channelized wheel path. Features ofthe settlements plates are illustrated in Fig. 4.4. To measure the thickness the top capwas removed and a measuring tape used to measure the depth to the bottom of thesettlement plate. These measurements were taken at certain intervals during traffickingand, knowing the initial elevation of the top cap and depth to the subgrade surface,settlement of the subgrade was calculated.Measurement of elongations in the geosynthetics during trafficking is importantin order to examine the influence of the base course layer thickness on mobilized strainin the geosynthetic. An array of fifteen marker studs was used to measure theelongation in the geotextiles. They were fixed transverse to the road alignment at tendifferent locations, see Fig. 4.3. The spacing between the studs is 15 cm underneathChapter Four - Description of Test Sections^ 74the influence zone of the tire load, and 30 cm in the middle of the sections and in theoutside anchorage zones. The exact initial spacing between the markers was measuredbefore placing the base course aggregate and the elongation was measured at certaininterval by trenching down to the geotextiles. A similar approach was used in the caseof the geogrid, but instead of attaching studs the initial length between every fourth ribjunction was measured prior to installation.Figure 4.4 Settlement plate - Details4.1.3 Construction of the Test SectionConstruction of the test sections started with removal of the topsoil vegetatedcover and excavation to the 0.06 m grade level referred in Section 3.4. Theconstruction equipment was not permitted to traffic the road subgrade duringconstruction, therefore all excavation and grading was done from the sides of the trialsection. The surface of the clayey silt subgrade was leveled to a elevation of 0.06 m+1- 1.5 cm.Chapter Four - Description of Test Sections^ 75The geosynthetics for the reinforced sections were placed directly on theprepared subgrade surface. Each geosynthetic was placed to overlap the adjacent testsections by 50 cm in the direction of anticipated fill placement. The geotextiles were4.5 m wide, and overlapping in the transverse direction therefore not needed. Thegeogrid, which was 4 m wide, was overlapped in transverse direction of about 1 m byplacing a additional 1.8 m wide strip along the length of the test section.The base course aggregate was also placed from the sides of the test sections inseveral lifts. The first base course layer was placed to a 15 cm loose thickness andcompacted using a small vibrating plate to a finished thickness of approximately12.5 cm. A small vibrating plate was used in all compaction work in order to minimizesubgrade disturbance. After the compaction of the first lift, the second lift was placedraising the finished thickness up to 25 cm, which is the minimum thickness in each testsection. Compaction measurements were taken to ensure the uniformity and theamount of compaction using a nuclear densometer. The specification was a minimumvalue of 90% of Standard Proctor. The 5% slope gradients were achieved by placingand compacting a wedge of gravel in two layers. The final surface was then leveled tothe target elevation after compaction to 96 % of Standard Proctor with a standarddeviation of 0.67 %. The final surface was measured using regular builders level andgraded to the target elevation +1- 5 mm using manual labor.The final stage of the construction was digging a small drainage ditch on bothsides of the test road. Since the elevation of the subgrade of the test area was lowerthan the surrounding environment, due to the excavation to grade level, ditches werenecessary to prevent water from draining into the test section from the surroundingsoil.Silt and ClayGravelFL.Core Hear0.150^10^1Grain size, in mmSand01020304050607080901000.011009080706050403020100Chapter Four - Description of Test Sections^ 764.2 BASE COURSE MATERIAL PROPERTIESThe base coarse material used in the road construction is defined as being well-graded, angular gravel with sand/sand with gravel, with coefficient of uniformity,Cu = 48, and the coefficient of curvature, Cz = 0.6. The grain size characteristic of theaggregate is shown in Fig. 4.5.The compaction characteristics of the base course aggregate were measuredemploying a standard Proctor test, according to ASTM D698, method C. TheGRAVEL/SAND has a standard Proctor maximum dry unit weight of 21.5 kN/m 3 atan optimum moisture content of 7.1 %. After placement and final compaction in-situ,the water content was measured using the nuclear densometer. At the in-situ dry unitweight of 20.7 kN/m3 the water content was 6.7 % with standard deviation of 1.15 %.Figure 4.5 Base course aggregate grain size distributionChapter Four - Description of Test Sections^ 774.3 GEOSYNTHETIC PROPERTIESProperties of the geotextiles used in the field trial are shown in Table 4.1. Thesevalues are based on manufacturers minimum average roll value data. Unfortunatelythere is no standardization of how properties of geosynthetics are reported and cantherefore be difficult to compare properties of geosynthetics from differentmanufacturers.Table 4.1 Properties of the geotextilesPolyfeltTS 600PolyfeltTS 700TexelGeo 9Property Units Value Value ValueStructure Non-wowen Non-woven Non-wovenPolymer Type Polypropylen Polypropylen PolypropylenThickness mm 1.8 2.3 1.8Mass Per Unit Area g/m2 204 268 310Apparent Opening Size mm 0.25 0.21 0.10Permittivity 1/sec 1.8 1.3 0.7Puncture kN 0.355 0.445 N/AMullen Burst kN/m2 1517 2207 2400Trapezoid Tear kN 0.311 0.38 N/AGrab Tensile/Elongation kN (%) 0.711 (50) 0.933 (50) N/AWide Width Strength/ElongationMD kN/m (%) 12.3 (95) 15.8 (95) 7.0 (10)XD kN/m (%) 10.5 (50) 14.0 (50) 7.0 (10)MD = machine (or roll) direction; XD = cross machine directionN/A = not availableChapter Four - Description of Test Sections^ 78It should be noted that the Texel Geo 9 is a reinforced non-woven geotextile,where a woven geotextile (Texpro 120) is used to reinforce a non-woven (Texel7609). Test data from several wide width tests on the Texel Geo 9, indicated anultimate load of 14 kN/m at about 21% strain.The properties of the Tensar BX 1100 geogrid are shown in Table 4.2. Thetensile strength of the geogrid is reported as a minimum specified value. Values ofwide width ultimate strength, based on manufacturers Q/A testing, show between 10 to70% higher values than reported, depending on the direction, where strength in thecross machine direction is considerably higher. Strains at failure are typically9 to 12%.Table 4.2 Properties of the geogridTensarBX 1100Property Unit ValueTensile- modulus @ 2% strain kN/m 204.0- ultimate strength kN/m 12.4Dimensional Stability- junction strength kN/m 11.2- junction efficiency % 90.0Geometry- aperture MD mm 25.0- aperture XD mm 33.0- rib thickness mm 0.76- junction thickness mm 2.80- open area % 70.0Material- Polypropylen % 98.0- Carbon Black % 0.54.34 m80.34 kN(8190 kg)2.03 m 2.15 m^0 32 m -^T T1.51mChapter Four - Description of Test Sections^ 794.4 LOADING VEHICLEThe load vehicle used to traffic the test sections was a single axle, dual tire truck,with a standard axle load (80.3 kN) and a tire inflation pressure of 620 kPa. A layoutof the wheel spacing and axle load of the test vehicle is shown in Fig. 4.6.Figure 4.6 Configuration of the loading vehicleThe truck was driven in forward and then in reverse over the entire length of thetest section at an average speed of 7 km/hour. Traffic was recorded in terms ofnumber of passes of a standard axle load. The front single-axle, single wheels of thetest vehicle has negligible effects on the road performance and was therefore discardedwhen reporting the number of passes of the load vehicle.Chapter 5Field Data5.1 INTRODUCTIONField testing took place between mid-December 1992 and late June 1993. Atotal of 500 passes were made with the loading vehicle during this time. In the earlystage of traffic loading only a few passes were made each day, and measurements weretaken frequently in order to get a better feel for the behavior of the reinforced soil-aggregate system and to determine the initial trend of base course and subgradedeformations. After eight passes, on the second day of traffic loading, the field trialwas postponed until the second week in February due to inclement weather conditions.Considerable effort was put in to keeping the general condition of the site and road asunchanged throughout the period of field trafficking. This was done by regular siteinspection and observations of the groundwater table in the drainage ditch parallel to80Fa' 16 m1 2 3 4 5 a 7 8 • 100.5^3.0 1.0^1.0^1.5 1.5^1.0^1.011 41111 .411.10Mill■MIONMIMEMIROMMISINIPIChapter Five - Field Data^ 81the test road, and covering the road surface with plastic sheets between the days oftrafficking.Measurements in each test section, where \"test section\" refers to one of the fivedifferent sections in the test road, are taken at pre-defined cross-sections in that testsection. Each test section is numbered in sequence from 1 to 5, with the unreinforcedbeing No.1 and the Polyfelt TS 600 No.5, as shown in Fig. 4.1.Location of marker studs for strain measurements = •and location of rut depth measurements =Scale 1:125Length In metersFigure 5.1 Cross-section numbers in each test sectionA total of ten cross-sections were monitored in every test section, see Fig. 5.1.Consequently, the test section number and cross-section number are combined andused as a reference system in the reporting of results: the first number refers to the testsection and second to the cross-section, hence location 2.3 is on Texel Geo 9,cross- section 3, where the base course thickness is 35 cm. Where \"left\" or \"right\" isused in describing the field observations, the definition of left assumes the observer islooking east from the unreinforced section (No.1) towards the reinforced sections(Nos.2 to 5).Chapter Five - Field Data^ 825.2 FIELD DATAMeasurements of surface profiles, ruts, subgrade profiles and settlements, andstrains in the geosynthetics were taken during vehicle trafficking. The data arepresented and described below.5.2.1 Surface ProfilesMeasurements of surface deformations across the width of the road were takenfrequently during the first 40 vehicle passes. The profile measurements were taken atevery cross-section in all test sections, giving a total of 50 measurements each time.Figures 5.2a, b and c show the development of surface deformation with increasingpasses at a base course thickness of 25, 35 and 50 cm, respectively. The threecross-section locations are 5, 3, and 1 respectively; they are selected for presentationbecause they also include the settlement plates. In the figures, elevation refers to thesurface of the base course and offset refers to the distance from the survey peg on theleft side of the road. The interception of the offset axis with the elevation is fixed atthe mean subgrade elevation of 0.06 m. A vertical:horizontal scale of 1.7:1 is used toclarify the nature of the deformations.The field observations of surface deformation, Fig. 5.2, indicate a clear symmetrybetween the dual wheels for each base course thickness and with increasing rut depth.Heave is localized between the wheels, and no significant vertical displacementoccurred at the longitudinal centerline. All profiles show a small difference indeformation between the inner and outer wheel path for each of the dual wheels, withthe outer wheel developing a slightly greater rut. It is felt this might be caused to someextent by the alignment of the front wheel of the vehicle with the outside wheel of therear assembly: however it is also recognized that the front axleSection 2 - 9- N 24N40-e_ N 700.8 ^0.5C^0.4 =.20.3▪ 0.2 -0.1 -Inked subgade surfacei(^.22Ui0.80.50.40.30.20.1 -0 4 453.50.5 11 40 2 31.5 2.5 450.5 3.50.80.5• 0.4.20.32LU^0.20.1Section 41111^11111Section 5 - 9- N 24-e- N • 700.5E.22Ui0.40.30.2453 3.5 40 0.5 10.8•0.11111111111Chapter Five - Field Data^ 83.5Ui0.5^1^1.5^2^2.5Offset distance (m)3^3.5^4^45Offset distance (m)Offset distance (m)Offset distance (m)Figure 5.2a Surface profiles at cross-sections 5, h = 25 cm, tests sections 1 to 51 42 31.5 ^^0.5 2.5 3.5•^0.5 1 3 3.5 4Chapter Five - Field Data^ 841I---••••••••••8.20.80.50.40.30.20.1Section 1^ N.0- N 8- 0- N.24N•40Mal subgrade surface0.6Initial stibgrade surface0^0.5 1 1.5 2 2.5Offset distance (m)3initial subsgade surface2 2.50 0.5 1 1.5 3 3.5 4 *^7Section 4- N 8-4,- N 400.60.5-e 0.40.3• 0.20.13.5^4^*sOffset distance (m)0.6Section 5^ / Initial subgrade surface11111111111.111^1111^- N 8- • • 24- 0- N • 401.5^2^2.5Offset distance (m)Figure 5.2b Surface profiles at cross-sections 3, h = 35 cm, tests sections 1 to 50 *«0.5▪ 0.4.20.3• 0.20.1.20.80.50.40.30.20.1Initial subgrade surface- N 03.5^4^4 5- N 8- 411-- N.40Offset distance (m)Section 20.5^1^1.5^2^2.5^3Offset distance (m)Chapter Five - Field Data^ 850.6 -0-0Section 10.50.4.20.3 -a0.20.1 -2 2.5^3^3.5^4^4 511- 1110.5 1.53 40.5 ^s3.51.5^2^2.5Section 3 nine subgrade surface11111111.220.30.2.220.30.20.11111111111111113 3.51.5^2^2.5Offset distance (m)0.60.50.4initial subgrade surfaceI^T^11111111111-AO-- N.8-AO-- N.40111/11^ [1111111Section 2- N 8.24- N 403.5^4^4 511 131.5^2^2.53 as 4 4 5.2.21.5^2^2.5subgrade surface- N- N • 40Section 4Initial subtrade surface0.20.1-11- N.8.24Section 540.5 4 51110.10.60.5• 0.40.30.20.10.60.50.4o 0.50.60.50.470.3onOffset distance (m)Figure 5.2c Surface profiles at, cross-sections 1, h = 50 cm, tests sections 1 to 5N.0- N 8subgrade surface-4,- N.40r IrrtrChapter Five - Field Data^ 86load is very small in comparison with the rear axle load.The 25 cm thick sections, Fig. 5.2a, show substantial heave on either side of thedual wheels. The heave is greatest on the outside of the dual wheels in most cases,which is in agreement with the larger rut in the outer wheel path. The behavior isattributed to interaction between the wheels, rather than an action of the geosynthetic,since this same trend is also apparent in the unreinforced section. The cross-sectionalarea of heave in these 25 cm thick sections appear to be less than the cross-sectionalarea of the ruts in all five sections: the response is attributed to compressibility of thebase course material in the early stage of trafficking, since the difference seems todiminish with increasing number of passes.The 35 cm thick sections, Fig. 5.2b, do not show any significant heave in thereinforced sections and a much smaller vertical displacement is observed in theunreinforced sections than in the 25 cm thick sections. The heave that did occur seemsto develop in the first few passes. The development of ruts without any significantheave, as in the case of the thicker base course reinforced and unreinforced sections, isattributed to some localized compaction of the base course. This same behavior isobserved in the 50 cm thick sections, Fig. 5.2c, where there is little heave in theunreinforced section and almost none in the reinforced sections.5.2.2 Rut DevelopmentRut measurements were conducted after every vehicle pass at cross-sections 1, 3and 6 (Fig. 5.1) in all test sections for the first 24 passes, and after every second pass inother cross-sections. Then measurements were taken after approximately every 10passes in all cross-sections up to 130 passes, every 20 passes up to 300 passes, andevery 50 passes thereafter.Chapter Five - Field Data^ 87Rut depth is defined as the difference between the initial average base courseelevation, before trafficking, and the elevation measured in the developed rut. Whendetermining the rut depth, four measurements were taken at a cross-section, two oneach dual wheel path, one in the inner wheel path and another in the outer wheel path.The measurements were taken at the lowest point in each wheel path. Since the testsections have two cross-sections of the same base course thickness, eight values wererecorded in total.The resolution of measurement is about ± 2 mm, which accounts for theelevation of the datum pegs which the reference string is attached, and themeasurement from the string down to the base course surface.The development of rut depth with number of passes for the five test sections isshown in Figs. 5.3 to 5.7. The upper figure for each section shows the relationshipbetween ruts and number of passes, for both the minimum base course thickness(25cm) and the maximum thickness (50cm). They illustrate the nature of the raw data:the data points collected from the two cross-sections of equal thickness in each testsection. Variations in the data are attributed to the two different locations of rutmeasurements for each base course thickness, and the slight difference in rut betweenthe outer and inner wheel path for the dual wheel assembly (see Fig. 5.2). Therepeatability of measurements is considered excellent, therefore the average value ofrut depth for each base course thickness is reported in the lower figure.Curves for the unreinforced section, Fig. 5.3, show two distinct general shapes.One defines a rapid development of rut at low passes followed by a nearly constant, butslightly increasing value with further trafficking. The other defines the same initialtrend of rapid development of rut at low passes, but is followed by continued rutting toa value of approximately 20 cm, which is considered a serviceability failure.Section 1 - Unreinforced5 0 IDE E^00 0gDEO2 PH ^ h=50cm• h=25cmI^'^I^I^I^I^'^I50^100^150^200^250^300^350^400^450^500Number of Passes, N•• , ■ ■ ■ ■ ■ ■•gie A • •0 o 0 1 0 ° °^0 0o 0 0 0 0 0 0■^^^^^^^ ^ ^^ ^ D 0000 0——————• •^■0^ h=50cm• h=40cm■ h=35 cmA h=30 cm• h=25cmI0^50^100I150^200^250^300I^'^I^I^'350^400^450^500Chapter Five - Field Data^ 88Figure 5.3a Rut depth versus number of passes - Section 1 raw data24222018E 16ft. 14cc•12e•>10< 86420Number of Passes, NFigure 5.3b Average rut depth versus number of passes - Unreinforced dataChapter Five - Field Data^ 89Section 2 - Texel Geo 924•• •2 ;•• •• • • 1I ••• ••• • •^••^• • •• ^1:1° Hr^He lep s eI• ••2220181814121086420^ h=50cm• h=25cmI^I^I^I^I•8•sQ80^50^100^150^200^250^300^350Number of Passes, N400^450^500Figure 5.4a Rut depth versus number of passes - Section 2 raw dataI^I^I^I^I^I^I^I^I24 —22 —20 •18••••••^A0004 21 slime!!!I^• • ••OA■■0A■^ h=50cmA •••h=40cmh=35cmh=30cmh=25cm161412100^FT ^I^1'1'10^50^100^150^200^250I^1'1'1'300^350^400^450^500Number of Passes, NFigure 5.4b Average rut depth versus number of passes - Texel Ego 9 dataE3524222018161412100•1:1^^■00-^ h=50cm• h = 40 cm▪ h=35 anA h=30cm• h=25cmChapter Five - Field Data^ 90^••^ _^so 8 : _• •...^•is $^ -•• •4, i • • ••• I • __• • • • •• •0-•.^• —0111.0 - 401• 6 0 508 'S .. 819 .li!hig °HH 11^ • h=50cm• h=25 anI^I^I^1^'^1^I^I'^I^I^I^I^150^100^150^200^250^300^350^400^450^500Number of Passes, NSection 3 - Polyfelt TS 700 -HUHFigure 5.5a Rut depth versus number of passes - Section 3 raw data2422 ••2018 • • A• ••• 17! ^▪ ! U !^!•• A A^im! ^ 000 ^1614121086420••A AA'^I^1 l '^r'^I'^1^'^I'^I0^50^100^150 200^250^300^350 400^450^500Number of Passes, NFigure 5.5b Average rut depth versus number of passes - Polyfelt TS 700 data• •0—Chapter Five - Field Data^ 91242220181814121086420- 2I .—-^•82•—-•••—- 44P.es,e* •— •—^• ^•-- es eS.—^•• li••• •OPOP Hg N NNi 'd IN ki_I'^I'^,^I'^,^,^I'^,0^50^100^150 200^250^300^350^400 450^500Number of Passes, NFigure 5.6a Rut depth versus number of passes - Section 4 raw dataI^I^I^I^I^I^I^I^ISection 4 - Tensar BX 11000— ^ h=50cm• h=25cmI^ h=50cm• h=40cm▪ h=35cmA h=30cm▪ h=25cm•••^ AAAA A• AA• AA• teigriaiiii ; P i Fi^ic Ol 602g242220181614121086420 I^I^I^I^I^I^I^I^I0^50^100^150^200^250^300^350^400^450^500Number of Passes, NFigure 5.6b Average rut depth versus number of passes - Tensar BX 1100 dataChapter Five - Field Data^ 92Et8re•242220••••• •• 11Section 5 - Polyfelt TS 600; 0 8•S e•••• 8••::!•IITHWPHHOki^Nkon^ h=50cm• h=25cm181614121086420 I^I^I^I^I^III0^50^100^150^200^250^300^350^400Number of Passes, NI^I450^500Figure 5.7a Rut depth versus number of passes - Section 5 raw data^ h=50cmO h=40cm•^▪ h=35cmA h=30cm• • h=25cm•242220184, „)^161412•10•86420A• a A^• ^A • A1 . 2I^Ici o oi• III°^o^■ M^^ ^ ^^mim°e& I; .0 20 00^o II •A°■^ RII ^0 0 0 0 0UEl-0^50^100^150 200^250^300^350^400^450^500Number of Passes, NFigure 5.7b Average rut depth versus number of passes - Polyfelt TS 600 dataChapter Five - Field Data^ 93A relationship is apparent between rut depth, number of passes and base coursethickness: smaller ruts are observed with increasing base course thickness for the samenumber of passes. The rut development in the unreinforced 25 cm thick section wasvery rapid and trafficking was stopped after 40 passes at an average rut depth of19.6 cm. Trafficking of the 30 cm thick cross-sections on either side of the 25 cmthick section was also discontinued, even though the average rut was only 13.6 cm:this was due to the need for repair of the 25 cm thick unreinforced cross-section(which was the first to fail). The repairing procedure was modified for later repairs inorder to keep the 30 cm thick cross-sections in the data base until a serviceabilityfailure was achieved. An estimation, based on the data obtained for the first 40 passesof the 30 cm thick unreinforced section, suggests a serviceability failure would havebeen achieved within the first 100 passes.Although the 35, 40 and 50 cm thick unreinforced sections did not reachserviceability failure, and show a very similar response, there is a noticeable differencein rut development. The final average rut depth after 500 vehicle passes for the 35, 40and 50 cm sections were 17.3, 14.4 and 11.3 cm, respectively.The five different test sections, unreinforced and reinforced, all show the sametwo general shapes of curves which characterize a response to trafficking which istermed either unstable or stable. Although there is a definite relationship between rut,number of passes and base course thickness for the reinforced sections, with lessrutting in the thicker subbase layers, it is not as consistent as that in the unreinforcedsection. The exceptions are in sections 4 and 5, see figure 5.6b and 5.7b, where the 40cm thick sections show less average rutting than the 50 cm thick cross-section after500 passes. This is attributed to compressibility of the base course material, and maychange with increasing number of passes in accordance with the observations forsection 3 between the 35 and 40 cm thick sections (Fig. 5.5b).Chapter Five - Field Data^ 94All of the 25 cm thick reinforced sections were loaded to a serviceability failure.Although there is a considerable difference in rut development between the reinforcedsections, they all demonstrate a significant improvement in behavior over the equivalentunreinforced section: test sections 2, 3, 4 and 5 failed at average rut depths of 21.6,21.6, 21.3 and 20.1 cm, after 300, 190, 140 and 150 passes, respectively.The 30 cm thick sections tended toward an unstable response, but did not reachthe condition of a serviceability failure after 500 passes. The curves show someperiodic variations in the relationship between average rut depth and the number ofpasses, see for example the 30 cm thick section in Fig. 5.7b. The behavior is attributedto the sequence of trafficking. It is felt some dissipation of excess pore pressuregenerated by vehicle loads took place during waiting periods between trafficking days.This would lead to an increase in undrained shear strength, and therefore a slightlygreater resistance to the onset of the next loads. This behavior was observed when theelapsed time between trafficking periods exceeded two weeks.The remaining reinforced test sections between 35 and 50 cm thick all show astable response, and an average rut depth at the end of the field trial that is less than theequivalent unreinforced section. A comparison of the reinforced section shows themagnitude of the rut depth is very similar for the 35, 40 and 50 cm thick sections,although there is some indication of slightly less rutting in the geogrid section. Itwould appear that the performance of these thicker sections, at 500 passes, isessentially independent of the type of geosynthetic. There is a difference in rut depthbetween the reinforced sections and the unreinforced section that is less significant forthe thicker sections: behavior of the 50 cm thick layers is almost identical for all of thetest sections.Chapter Five - Field Data^ 955.2.3 Subgrade Profiles and SettlementsDeformations of the subgrade surface were measured at certain stagesthroughout the trial by trenching down to the surface of the geosynthetics andrecording vertical displacement with respect to a horizontal datum. Most of thetrenches extended over half the width of the road to minimize any disturbance of thesystem. They were about 30 cm wide, and after taking the measurements they werebackfilled, compacted and graded to the original profile. No noticeable difference inbehavior was observed prior to and after trenching.Subgrade profiles are illustrated in Figs. 5.8 to 5.11. All the 25 cm thickcross-sections were trenched after 80 passes, shown in the upper part of the figures.Consider the 25 cm thick base course layer. Little significant deformationsoccurred in section 2.5 (Fig. 5.8) after 80 passes. In contrast, other three reinforcedsections show more distinct deformed subgrade profile and a tear was observed in thegeogrid in the left hand side wheel path (Fig. 5.10). The geotextiles developed asmoother profile than the geogrid, which exhibits a noticeable difference between theinner and outer wheel path. During trenching of the 25 cm thick geogrid reinforcedsection, it was observed that the subgrade soil had squeezed up through the aperture ofthe grid openings about 2 cm: however this was not observed in the 35 cm thickgeogrid section after 500 passes where the base course aggregate, the geogrid, and thesubgrade, were well interlocked at the interface.The deformations in the 25 cm thick cross-sections generally show a fairagreement between the volumetric displacement in the wheel path and the volumetricheave on either side of the path after 80 passes. One exception is section 4.6, shown inFig. 5.10, but that is to some extent attributed to the squeezing of the subgradethrough the geogrid apertures.Initial averagetexlile/subgradeInterfaceN - 80Chapter Five - Field Data^ 960.3^II^tills^till^,0.2 = Texel Geo 9^X-Section 2.5 ; h = 25 cmE^0.1 _niI■111%,I,^tUs0 0.5 1.5^2^2.5 3 3.5 4 45 Offset distance (m)..\"..\"[...■^ I^i^1 I 1 \" 0.30.20.1^.2^0^-2^-0.1-0.2-0.3Texel Geo 9 X-Section 2.3 ; h = 35 cmInitial averagetextile/subgradeinterfaceN - 5000 0.5 1.5^2^2.5Offset distance (m)3 3.5 4 45Figure 5.8 Subgrade profiles in test sections 2.5 and 2.3 - Texel Ego 90.30.2E^0.1.2^0A^-0.1-0.2-0.3Polyfelt TS 700^ X-Section 3.5 ; h = 25 cm_^...^,.Adaibtak,^\\_.......ari----- -----^N = 40- -- - N = 80N =160- ._ --initial averagetextile/subgradeinterface^ A.____- observed tear-1.5 450.30.2E^0.1.2^0-0.1-0.2 --0.3^^0.5.,.,t..,Polyfelt TS 7002^2.5^3Off, set distance (m)11111111111111111111 X-Section 3.3 ; h = 35 cmInitial averagetextie/subgradeinterface^ N - 50o3.5^40 0.5^1 1.5^2^2.5Offset distance (m)3^3.5 4 45Figure 5.9 Subgrade profiles in test sections 3.5 and 3.3 - Polyfelt TS 700Offset distance (m)Tensar BX 1100^X-Section 4.3 ; h = 35 cmInItIal averagetexthe/subgradeInterface111,11111.1111111[1 .0.30.20.1C^-.2 0• -.1-0.2^-0.3^^N - 5002.5 3 3.5^4 45451.5^2^2.5^3^3.5^4-0.30.30.2E 0.1.2•^0-2^-0.1 -w-0.2 -111111110.5I^I^IPolyfek TS 6001.5^2Offset distance Cm)Initial averagetecele/subgradeInterfaceN - 160IX-Section 5.3 ; h = 35 cmN 5001111 , 11111-1111. 11111111111IITY-0.30.511111/11^11111,1f11Chapter Five - Field Data^ 970.3^'^IE Tensar BX 11000.2 =I^I^I X-Section 4.6 ; h = 25 cmE 0.1c.2^0 --WA -0.1-0.2-0.3 ^0observed tear /24 Ink's! averagelextile/wbgrsdeInterfaceN - 800.5^1^1.5 2^2.5 3 3.5 4 450 0.5^1^1.5^2^2.5^3^3.5^4^45Offset distance (m)Figure 5.10 Subgrade profiles in test sections 4.6 and 4.3 - Tensar BX 1100Offset distance (m)Figure 5.11 Subgrade profiles in test sections 5.5 and 5.3 - Polyfelt TS 600Chapter Five - Field Data^ 98Progressive observations for cross-sections 3.5 and 5.5, see Figs. 5.9 and 5.11respectively, show a common turning point at the subgrade surface marked TP in thefigures. The turning point separates rutting in the wheel path from heave outside it,and is an indication of that the load spread angle through the base course layer wasincreasing with increased rut depth. The symmetry of the deformed profiles isattributed to an equal anchorage resistance on either side of the turning point. Itshould be noted that cross-section 3.5 was repaired after 190 passes when it hadreached a state of serviceability failure, however it was later excavated after 280 passesand the information included in Fig. 5.9.The 35 cm thick cross-sections were excavated after 500 passes, but show nodistinct pattern of deformation and no significant difference in relative performance isapparent.A further evaluation of subgrade displacement can be made from the settlementplate data reported in Figs. 5.12 to 5.16. The plates were located underneath the leftdual wheel path, and consequently do not necessarily indicate the maximum verticalsubgrade deformation. However, the magnitudes of displacement measured from thesettlement plates and from the subgrade profiles are in good agreement. Oneinteresting common trend in the curves is that during the first ten passes there was arapid increase in vertical displacements to about 3 cm. The system appears to stabilizeafter this initial settlement, and the increase of vertical displacement with number ofpasses is much slower. There is however a noticeable difference in behavior betweenthe three base course thicknesses, with the exception of the very stiff Texel geotextile.There is also a difference in behavior between the test sections, at 25 cm and 35 cmthickness, where the smaller settlements of the Texel and Tensar sections (Figs. 5.13and 5.15) are attributed to the greater stiffness of these materials. No significantdifference in behavior is observed in the 50 cm thick section.3 L:45eFco^7 -I:8 =9Base Course Thickness-8- h =50 cm ; X-Sec.1.1-e- h = 35 cm ; X-Sec.1.3h = 25 cm ; X-Sec.1.1310 I^I^I I200^2500^50^100^150 300^350^400^450^500Section 1 - UnreinforcedBase Course Thickness-8- h = 50 an ; X-Sec.2.1-G- h = 35 an ; X-Sec.2.3—A— h =25 an ; X-Sec.2.60T^I^I^I^I^I'^I^I^I^I'^I50^100^150^200^250^300^350^400^450^500Number of Passes, N12E 345213.rn 78Chapter Five - Field Data^ 99Number of Passes, NFigure 5.12 Subgrade settlement - Section 1Figure 5.13 Subgrade settlement - Section 2Base Course Thickness—B— h = 50 an ; X-Sec.3.1—G— h = 35 an ; X-Sec.3.3—A— h = 25 an ; X-Sec.3.610O 50^100^150^200^250^300^350^400^450^500Chapter Five - Field Data^ 100Number of Passes, NFigure 5.14 Subgrade settlement - Section 3I^I^I^I^I^11111 Section 4 - Tensar BX 1100 1^ECBase Course Thickness- h = 50 an ; X-Sec.4.1-9- h =35 an ; X-Sec.4.3h = 25 an ; X-Sec.4.6I^'^I-^1^I^I^'^I^1^I 1^1^1^1^1O 50^100^150^200^250^300^350^400^450^500Number of Passes, NFigure 5.15 Subgrade settlement - Section 401101^I^1^I^I^11^1^1^1^1^I Section 5 - Polyfelt TS 600 -I0I^I^I^I^1^1^1^I^1^I^I0^50^100^150^200^250^300^350^400^450^50010^ECBase Course Thickness-9- h =50 cm ; X-Sec.5.1-G- h =35 cm ; X-Sec.5.3- h = 25 cm ; X-Sec.5.69Chapter Five - Field Data^ 101Number of Passes, NFigure 5.16 Subgrade settlement - Section 55.2.4 Geosynthetic Strain MeasurementsElongation of the geosynthetics was measured at the time of trenching. All ofthe 25 cm thick sections were measured at the location of cross-sections No.5 andadditionally at cross-section No.6 in the Polyfelt test section to determine therepeatability of behavior. Measurements were also taken in the 30 cm thick Texel andPolyfelt TS 600 sections after 280 passes. Strain measurements for 35 cm thicksections were taken after 500 passes for all test sections in cross-sections No.3. Theresolution of measurements is ± 0.5%The 25 cm thick layer of section 2, Fig. 5.17, shows a distinct increase in tensilestrain in the fabric with increased number of passes. However, the strain magnitudeafter 280 passes is not significantly larger than that after 80 passes, even though the--------- N = 80^ N - 280X-Section 2.5 ; h 25 cm-8-6-4-20246810121416Texel Geo 9-8-6-4Texel Geo 9-220c 46N - 500(4 8101214160 3 4 450.5 3.51.5^2^2.5X-Section 2.3 ; h = 35 cmChapter Five - Field Data^ 102difference in rut depth is considerable: this section reached the serviceability failurecriterion after 300 passes. No significant strains were observed in the 30 and 35 cmthick layers.0^0.5^1.5^2^2.5^3^3.5^4^45Offset distance (m)8 ^t43 - Texel Geo 9^ X-Section 2.7 ; h = 30 cm =-4-202 74 -68 710 -12 -14 -16 N = 280I05^1^1.5^2^2.5^3^3.5^4^4 5Offset distance (m)Offset distance (m)Figure 5.17 Geotextile strain measurements - Section 2Strain measurements from the 25 cm base course thicknesses of section 3, atvarious numbers of passes, and from the 35 cm thick section after 500 passes, areshown in Fig. 5.18. There is a definite indication of symmetry in the 25 cm thicksection, and a significant increase in tensile strain in the textile with increased numberPolyfelt TS 700-8-6-4-20246810121416X-Section 3.3 ; h = 35 cm N - 60oChapter Five - Field Data^ 103of passes. The strain profiles show very close agreement with the observed subgradedeformation profiles discussed previously. After 500 passes on the 35 cm thick layer, alarge compressive strain is recorded when no significant tensile strain is developed(Fig. 5.18): this compressive strain is attributed to folding during placement of thegeotextile. Similarly no significant strain is detected in the 35 cm tick layer of sectionNo.4 or No.5 after 500 passes, see Fig. 5.19 and 5.20.-,X-Sections 3.5 & 3.6 ; h = 25 cm....^„,..^.....^.^,.-- N = 80.........__.-. ^N=160X-See.3.6 - .^-^N = 280Polyfelt TS 700 N - 280-8-6-4-202468101214160 0.5 1.5 2 2.5 3^3.5 4 45Offset distance (m)0^0.5^1^1.5^2^2.5^3^3.5^4^45Offset distance (m)Figure 5.18 Geotextile strain measurements - Section 3I^I^I X-Section 4.3 ; h = 35 cmTensar BX 1100 N - 500-2024681012141605 2.5^3^3.5Offset distance (m)Figure 5.19 Geogrid strain measurements - Section 4I1.5^2 4^4505 4543.53-4-202•c 46to 810121416-8-6Polyfelt TS 6001^1^I^I^1^i^I^l^i^i^1111111111^itIllX-Section 5.3 ; h = 35 cmN 500Chapter Five - Field Data^ 1048-4 Polyfelt TS 600.........^.-202 -c^46(1)^810 -12 -1416 ^towedX-Sec.5.5 ; N =160^ X-Sec.5.6 ; N = 160X-Sec.5.5 ; N = 12-^X-Sec 5 5 • al =80 h = 25 cmI^I0 0.5^1^1.5^2^2.5^3^3.5^4^4 5Offset distance (m)05^1.5^2^2.5^3^3.5^4^45Offset distance (m)Figure 5.20 Geotextile strain measurements - Section 5There is strong indication that the distribution of the strain underneath the loadedarea in the wheel path is highly non-linear. There is also evidence of a significantdifference in tensile strain between the inner and outer wheel of the two dual-wheelassemblies, even at relatively low number of vehicle passes. The pattern is mostevident in the thinner sections.Chapter Five - Field Data^ 105A summary of the maximum tensile strain in each-cross-section is given inTable 5.1 together with calculated value of average strain in the geosyntheticunderneath the loaded area.Table 5.1 Summary of geosynthetic strain measurementsBase CourseThicknessGeosyntheticNumber of Passes, N12 40 80 160 280 500C (%)max^(avrg)s (%)max^(avig.)s (%)max^(avrg.)s (%)max^(avrg)6max(%)(avrg)s (%)max^(mg.)h = 25 cmX-Sec.2.5X-Sec.3.5X.Sec.3.6X-Sec.5.5X-Sec.5.6Texel Geo 9Polyfelt TS 700Polyfelt TS 700Polyfelt TS 600Polyfelt TS 6002.0 (0.3)2.0 (0.9)2.63.37.4(0.9)(2.1)(5.2)9.516.012.0(6.8)(12.1)(10.0)6.613.915.2(4.4)(8.5)(9.4)h = 30 cmX-Sec.2.7X-Sec.5.4Texel Geo 9Polyfelt TS 6004.013.8(3.5)(10.7)h= 35 cmX-Sec.2.3X-Sec.3.3X-Sec.4.3X-Sec.5.3Texel Geo 9Polyfelt TS 700Tensar BX 1100Polyfelt TS 6003.32.61.34.0(1.5)(1.0)(0.6)(1.1)There is again clear indication of increasing strain with increased number ofvehicle passes. Also evident is a significant difference in the magnitude of tensile strainbetween the four different geosynthetics, at the base course thicknesses of 25 and 30Chapter Five - Field Data^ 106cm, which is attributed to the variation of stiffness of each material. The least stiffgeotextile, Polyfelt TS 600, exhibits the largest strain of 16%, and the stiffestgeotextile, Texel Geo 9, exhibits the smallest strains. Very small strains are observedin the 35 cm thick layer: no significant difference is observed between the three textiles,but the geogrid developed significantly less strain after the same number of passes.Chapter 6Interpretation of Field Data6.1 INTRODUCTIONAn interpretation of the field observations is presented here which addresses theinfluence of the geosynthetics, the performance of the unreinforced test section and theperformance of the reinforced test sections. The intent is to verify some of theassumptions made in design methods, and compare results with the approach of Giroudand Noiray (1981).6.2 INFLUENCE OF THE GEOSYNTHETICSComparison of the average measured rut depth versus number of passes in thetest sections for each base course thickness (Figs. 6.1 to 6.5) indicates all of the 25 cm107Chapter Six - Interpretation of Field Data^ 108thick sections are unstable, where unstable is defined as a rapidly increasing rut withincreasing number of passes. However, there is significant difference in behaviorbetween the unreinforced and the reinforced sections, particularly for the Texel Geo 9geotextile. Examination of the number of passes at the same average rut depth, showsthe Texel Geo 9 takes between 8 to 10 times more vehicle passes than the unreinforcedsection; the other reinforced sections take between 3 to 4 times more passes than theunreinforced section.The varying nature of the improved response of the 25 cm thick section isattributed to a contribution from the geosynthetics which differs for each material. TheGeo 9 material shows the greatest improvement, even at a very low numbers of passes.The other three geosynthetics exhibit a very similar trend up to about 50 to 75 passes,after which some difference in performance is observed. The response of thegeotextiles is consistent with their tensile strength and is attributed to their stiffness.The biaxial geogrid, BX 1100, shows similar response to the TS 600 and 700geotextiles up to about 50 passes, at which point the rut development accelerated. It isfelt the separation function is of great importance when the base course thickness isrelatively small. The initial performance of the geogrid is attributed to its higherstiffness and an ability to resist lateral spread of the base course aggregate by interlock.As subgrade intrusion occurs, the interlock is reduced and with it the mechanism oflateral restraint.The 30 cm thick sections are also categorized as unstable since there is fairlyrapid rut development with increasing number of passes, see Fig. 6.2. Comparisonshows that the unreinforced section gives a slightly better performance than the 25 cmsection, but the reinforced sections improve significantly. Although the relationshipbetween geosynthetic stiffness and performance is less obvious, the stiffest geotextilestill gives the best performance. It is felt that separation is still of some importance, butChapter Six - Interpretation of Field Data^ 109less so than for thinner base course. This is apparent in the relative performance of thegeogrid, which is better than two of the three geotextiles. Interestingly the test sectionwith the stronger, TS 700, geotextile developed a poorer response to trafficking at lessthan 150 passes than the Polyfelt TS 600. It should be noted that after about 120passes, a tear occurred in Polyfelt TS 600 fabric in the left wheel path. The tearoccurred where before installation in-uniformity in the fabric thickness was observed.This explains why Polyfelt TS 600 gives a different response after 120 passes.Figures 6.3, 6.4 and 6.5 show average rut depths with number of passes for thebase course thicknesses of 35, 40 and 50 cm, respectively. All of these test sectionsdeveloped a stable response characterized by a rut depth that is almost constant orslightly increasing with increasing number of passes. Although there is a significantdifference in performance of the unreinforced and reinforced sections in the 35 cmthick section, there is no significant difference between the geotextile reinforcedsections, and a very small improvement in the geogrid reinforced section over thosewith geotextiles. This behavior appears to be independent of stiffness, and therefore isattributed to a better interlock between the base course layer and the geogrid, whichtends to reduce lateral displacements of the aggregate and therefore preventingsubgrade intrusion.The difference between the unreinforced and the reinforced sections reduces withincreasing base course thickness, and little significant difference is observed in the 50cm thick section, see Figs. 6.4 and 6.5. Application of linear regression to theobservations suggests there would be no difference between them for a base coursethickness of 60 cm at 500 passes.• •••• •• Unreinforced• Texel Geo 9^ Pdyfelt TS 700▪ Tensor 8X 1100• PdYfek TS 600'^I^1200^250^300^350^400^450^500Number of Passes, N2422201800O642Chapter Six - Interpretation of Field Data^ 11024220 0^ •Base course thickness = 25 cm■o0■ooO ■ ooo ■o08^li g °8 •ino8)^so^ip.0 0 I • •*M^• •••••1 soO^ ••0^50^100^1502018O• 168 145• 12• 10•< 86420Figure 6.1 Comparison of average rut depth versus number of passes : h = 25 cmI^I^I^I^I^I^I^I^IBase course thickness = 30 cm09 I •0^■^•■^•■Pp • ■ ■^••^o^^ 0^• •„ 0 00—fl mi^••^••I 1116.ol:Unrelnforced■• Tecel Geo 9Polyfelt TS 700Tenser EIX 1100Pdyielt TS 600•0 I^I^I^I'^I^1'1'1'1'0^50^100^150^200^250^300^350^400^450^500Number of Passes, NFigure 6.2 Comparison of average rut depth versus number of passes : h = 30 cm10840 O▪ g0■ ■ ■^f(I t■Unreinfcrotte• Texel Geo 9^ PdyfsS TS 700▪ Tenser BX 1100• Pc4yfelt TS 600o6 9 0 0O0.WooD°W. 'le °'0000 00 0 0 0 0 0i:le^oo o ij• D^()0 0 0 00■ ■ ■ gi ■ ■24222018 —181412 —O ooo I: I I:I•••O 00■ ■■Base course thickness = 35 cmChapter Six - Interpretation of Field Data^ 1111^1^I^1 0 1'1'1'1'1^I^I'^1^I0^50^100 150^200^250^300^350^400^450^500Number of Passes, NFigure 6.3 Comparison of average rut depth versus number of passes : h = 35 cm124 —• Unreinforced• Teas! Geo 9^ Polyfelt TS 700• Tenser BX 1100• Polyfen TS 600Base course thickness = 40 cm2220188o01814121088000 00000 0if _ 00 00^^0 El lii .•••••M00;: 000O 0 0 0 0 0 o 0 o O^0^0^0^ ^ El^1 a• ^ .0 ^ ^ .0 .0 .0 0. • • •• • •^o ^0 r■0 0 so•..•..••U ■^■ ■^■o111-• ■ ■ ■ ■2 —1^I^1'1'1'1^1^1^10^50^100 150^200^250^300^350^400^450^500Number of Passes, NFigure 6.4 Comparison of average rut depth versus number of passes : h = 40 cm0Chapter Six - Interpretation of Field Data^ 11224222018Ut31814W 12010>< 86420- c<><><><><><><>- \"><> • MA @el SS al IP ■^•^•^•^0-— ■ ■ ■ in--11). 122::::: ■ •• ■ ■ ■ ■ ■ ■ —-10I^I^I^I50^100I^I^I^I I^I^I^I^I^'150^200^250^300^350^400^450^500• Unnanforcode Taal Goo 9^ Pdyfelt TS 700• Taw ED( 1100O Walt TS 600Base course thickness = 50 cmNumber of Passes, NFigure 6.5 Comparison of average rut depth versus number of passes : h = 50 cm00 0 0 0 * 0 0The relationship between subgrade settlement and number of passes, Fig. 6.6,was obtained from the settlement plates (see Fig. 4.3). Since three settlement plateswere installed in each section, only one measurement, and not an average, is shown inthe figure.There is a significant difference between subgrade settlements of the unreinforcedand reinforced systems below the 25 cm thick section, which occurs at the onset oftrafficking. The response is attributed to a tensioned-membrane effect. The reinforcedsections also behave differently, with sections 4.6 and 5.6 showing a rapid settlementthat is consistent with observations of surface rutting. It is concluded that, for this thinbase course layer, both stiffness of the geosynthetic and adequate separation betweenthe subgrade and base course layer are very important.Base course thickness = 25 cm2O 3X-See. 1.6X-Sec. 2.6-9- X-Sec. 3.6- X-Sec. 4.6- X-See. 5.6104• 57• 89Number of Passes, N123••• 452'^789—e— XSsc. 1.1- X-Sec. 2.1- EF x-sec. 3.1- X-Sec. 4.1—e— ksec. 5.110Number of Passes, NFigure 6.6 Subgrade settlement versus number of passes : h = 25, 35 & 50 cmChapter Six - Interpretation of Field Data^ 1130^50^100^150^200^250^300 350^400^450^500Base course thickness = 25 cmX-Sec. 1.6-0- X-Sec. 2.6-8- X-Sec. 3.6- X-Sec. 4.6-0-- X-Sec. 5.6012345.c7Base course thickness = 50 cm1234I^I^I200^250^300^350^400^450^500Number of Passes, N1150Chapter Six - Interpretation of Field Data^ 114Number of Passes, NFigure 6.7 Changes in base course thickness with number of passesChapter Six - Interpretation of Field Data^ 115Subgrade displacements below the 35 and 50 cm thick reinforced sections at theonset of trafficking are larger than those for the 25 cm thick section: The comparativeresponse is consistent with a tensioned-membrane effect that is negligible, and a similarresponse in the unreinforced and reinforced systems.The difference (Oh) between the measured rut depth and the measured subgradesettlement from the settlement plates, see Fig 6.7, is established from the same wheelpath and cross section where the settlement plates are located. A horizontal lineindicates the base course layer is moving with the subgrade, and no change in basecourse thickness occurs beneath the vehicles wheels. Hence volumetric changes in thebase course layer can be quantified, assuming no lateral spread. Significant ruttingoccurs during the first 8 to 10 passes of trafficking, and is accompanied by about 8 %compaction (change in thickness) in the reinforced sections and between 12 and 17 %in the unreinforced section. After about 75 passes, incremental displacements of thebase course and the subgrade are equal in the 35 and 50 cm thick sections. At thispoint, compaction of the reinforced base course layer is approximately 12 % andbetween 17 and 22 % in the unreinforced section. It is possible that some of thischange in layer thickness is due to lateral spread of the gravel beneath the wheels, but itis likely not significant.Unifying the observation of average rut depth, subgrade settlement and change inbase course thickness versus number of passes, three general stages of response totrafficking are identified in the 35 and 50 cm sections. Firstly, there is rapid rutdevelopment with number of passes, up to 10 passes, at which point approximately70 % of the total base course layer compaction has occurred and all the initial subgradesettlement is complete. Secondly, there is a significant increase in rut developmentI^' 1^1^1^1^135 40 45 5025^30ISection 1 - Unreinforced024 —22 —64 —2 —Chapter Six - Interpretation of Field Data^ 116combined with a less significant further subgrade settlement up to around 75 passes, atwhich point full compaction of the base course layer is obtained. Thirdly, after 75passes and at full compaction, vertical displacements in the base course layer and thesubgrade are equal.The relationship between rut depth and base course thickness is shown inFigs. 6.8 to 6.12. The data at few vehicle passes, say N equals 10 to 15, indicate theresponse of any section is independent of the base course thickness for all sectionsexcept perhaps the Geo 9 (Fig. 6.9).A very stable response is observed in the reinforced sections when the basecourse thickness is larger than 40 cm. In contrast, at a base course thickness of 35 cmor less, the rut development increases with decreasing base course thickness and thesystem is considered to be unstable. It is however not significant, but might becategorized as a serviceability failure when the ruts are deeper than 10 cm.Base Course Thickness (cm)Figure 6.8 Rut depth versus base course thickness - Section 1t I^ilt^II^II^1^II^I^I^I2 —o 1^T 1 1 1 125 30 3524 —22201818i' 148 12re• 10Section 2 - Texel Geo 9 —0— N —10—40-- N = 20-EH N 40—.— N = 800^ N — 160AP— N = 300—A^ N . 50040^45^501I^I^I^I^I^1^1^1^1^I^1^1^1^I^11 i iT420 1^1^F 11^T^I24221818i 14ta 12'5ft^10Section 3 - Polyfelt TS 700 N — 10—4,-- N = 20-El- N 40--II-- N = aoN —180—IIII-- N= 300—A— N- 500-El—40-—025 30^35^40 45^5086Chapter Six - Interpretation of Field Data^ 117Base Course Thickness (cm)Figure 6.9 Rut depth versus base course thickness - Section 2Base Course Thickness (cm)Figure 6.10 Rut depth versus base course thickness - Section 324 Section 4 - Tensar BX 1100 -0-- N - 10—4- N - 20- N - 40N = 800^ N - 110- N=300- N - 50022201881614121086 ^B^424 Section 5 - Polyfelt TS 600 N - 10N - 20N - 40- N = 800^ N - 150N 300-A^ N = 50020■ 1^I^It 1^I^130 35^40^45 502560Chapter Six - Interpretation of Field Data^ 1181^I^I^1 I 2 — 0 I 1^125^30^35 40^45^50Base Course Thickness (cm)Figure 6.11 Rut depth versus base course thickness - Section 4Base Course Thickness (cm)Figure 6.12 Rut depth versus base course thickness - Section 5Chapter Six - Interpretation of Field Data^ 119Very common response is observed in test sections having base course thicknessgreater than 35 cm. In the 35 cm thick sections, rut development is approximately at aproportional rate to doubled number of passes, whereas in the 40 and 50 cm thicksections a more stable response is observed, associated with a relatively steady statepoint in the 50 cm sections. As previously mentioned, no difference in rut between testsections would be anticipated for sections 60 cm or greater at 500 passes, and thesteady state point would be very clear after certain number of passes, when compactionof the base course layer has taken place and the initial subgrade settlement.It should be noted that this comparison is all related to 500 passes and if thenumber increases the stable response would move towards the thicker sections but asimilar response as previously showed is anticipated.6.3 UNREINFORCED PERFORMANCEA curve that establishes the minimum design thickness of an unreinforced basecourse layer for different numbers of vehicle passes was proposed by Hammitt (1970),based on a full scale field trial. The equation that described the curve was thenmodified by Giroud and Noiray (1981) to include various axle loads and rut depths.Three sets of curves, for a standard axle load, using Giroud and Noiray's proposedequation are shown in Fig. 6.13. Three different values of undrained shear strengthwere used that bound the strength of the subgrade at the site of this research study, andtwo values of rut depth were used that bound the displacements mobilized by trafficloading in this study. The curves of base course thickness versus number of passes arevery sensitive to the thickness and a small increment in base course thickness changessignificantly the number of passes.O^•O ■O^■Field Data(Austin and Coleman, 1993)• rut =5 an• rut =10 an10UBC Field DataO rut =5 cmI. rut =10 an• rut =15 cm100Chapter Six - Interpretation of Field Data^ 120The field data from this study are plotted together with the Giroud and Noiraycurves. The independence of rut depth and base course thickness is shown clearly forthe case of a 5 cm rut depth and few vehicle passes (N<10) in this study. Withincreasing rut depth, when the dependence on the base course thickness is mobilized(at a rut depth of 15 cm), there is reasonable agreement with the empirical equation:although the magnitude is slightly different the trend is the same. It would appear theempirical equation does, however, predict more vehicle passes at a given rut depth thanobserved from the field trial.Number of Passes, NFigure 6.13 Unreinforced data comparisonHowever, it should be noted that the original equation proposed byHammitt (1970) was derived by employing the method of least squares fit to datapoints that had considerable scatter. Furthermore, it does not consider any additionalstrength contribution from a base course layer with a CBR value greater than 11, e.g. abase course layer having a CBR value greater than 11 is considered to be equivalent.Chapter Six - Interpretation of Field Data^ 121While it is likely that the difference between the field data and theoretical curves is dueto properties of the base course material, the compaction of the base course prior totrafficking and the derivation of the initial equation, the theoretical curves have alsobeen extrapolated for rut depths other than those addressed by Hammitt.Results from the field trial described by Austin and Coleman (1993) are includedin Fig. 6.13. The subgrade strength was reported as a CBR of less than one, and thevehicle loading was comparable to this study. The trial comprised nine 6 m long by6 m wide test sections, each approximately 20 cm thick; three were unreinforced, andsix were reinforced with different geosynthetics. Measurements were only taken at onelocation within each test section. Reasonable agreement is observed with the empiricalcurves for more than ten passes, but the rut development in the three unreinforcedcontrol sections is very inconsistent and results indicate that base course thickness ofapproximately 20 cm develops a 5 cm rut after 2 to 18 passes. This is attributed tovariation in both subgrade and base course strength.6.4 REINFORCED PERFORMANCEThe performance of the reinforced sections is compared with Giroud andNoiray's design procedure, see Fig. 6.14. The input parameters used in developing thischart, following the Giroud and Noiray semi-theoretical design procedure, were basedon the material properties and the loading vehicle used in the field trial: c u = 40 kPa;P = 80.3 kN; pc = 620 kPa; e = 1.83 m; and K = 15 kN/m. Once again it is apparentthat development of the 5 cm rut is relatively independent of the base course thickness,likely as a result of initial compaction of the base course layer. For the 10 and 15 cmrut depths, the trend between base course thickness and number of passes for the fieldChapter Six - Interpretation of Field Data^ 122data shows good agreement with the semi-theoretical design chart. However, theabsolute magnitude of the predicted ruts is slightly different, causing the designapproach to over-predict the performance of the field trial. This difference is in partattributed to the compression of the base course, which is not taken into account in thedesign procedure.50^ \"op'^ • 45 740 —E 35 2.• —If)• 3025 —gO 2015 —10 —5Field DataA • Section 2; rut = 5 cmA^Section 2; rut =10 anA^Section 2; rut =15 anO Section 3; rut = 5 cm• Section 3; rut = 10 an• Section 3; rut =15 an• Section 4; rut = 5 cm▪ Section 4; rut = 10 an▪ Section 4; rut = 15 anO Salim 5; rut = 5 cm▪ Section 5; rut = 10 an• Section 5; rut = 15 anO ^ 0^A0^^ 0 A EDco^a^4^IS A0 0 o OD ■ 0- Rut7 (an)51015Giroud & Noirey procedureI^I^I^I^III10^100^500Number of Passes, NFigure 6.14 Reinforced data comparisonIt is also attributed to how Giroud and Noiray treat the unreinforced sections, based onthe data from Hammitt (1970), since the quasi-static analysis performed for thereinforced case does not account directly for the number of passes but takes them intoaccount through a manipulation of the unreinforced data from Hammitt. Examinationof the field data in Fig. 6.14 for the geogrid section (section 4) shows it is the onlysection that experience a greater increase of number of passes, with base course layerthickness, for the same rut depth compared with the Giroud and Noiray's procedure.Based on the fact that the differences in magnitude are not directly proportional for the0Chapter Six - Interpretation of Field Data^ 123unreinforced and the reinforced data, and that the subgrade settlements are not directlyproportional to the base course settlements, it is concluded that Giroud and Noiray'sprocedure is over-predicting the membrane effect on the overall performance of thereinforced sections.6.5 LOAD DISTRIBUTION ANGLEThe load distribution angle is of importance, when designing for a minimum basecourse thickness employing a pyramidal load distribution, because it is used todetermine the vertical stress acting on the subgrade.50E840• Tex! Geo 9^ Pdyfelt TS 700▪ Tamar BX 11000 IN:4*S TS 600 30 0 m ^•20■ ^0100^I^I^I^I^I '^I^I^1^I I^1^1^1^I^I^I^1^1^I^I^I^1^1^I^I^1^1^I'^1^1^1^I^1^1^1^1^I '^0^5^10^15^20^25^30^35^40^45Load Distribution Angle, a (deg.\"Figure 6.15 Load distribution angle versus base course thicknessValues of mobilized load distribution angle were deduced from comparison ofthe surface profiles (Fig. 5.2), the subgrade profiles (Figs. 5.8 to 5.11), and the changein base course thickness (Fig. 6.7), and are reported in Fig. 6.15. The load distributionChapter Six - Interpretation of Field Data^ 124angles were determined where data of subgrade profiles were available and of areasonable shape, and therefore only a few load distribution angles are obtained. It canbe seen, Fig. 6.15, that the load distribution angle varies with depth and between testsections. The range is between 17° and 40°, which compares reasonably well with thatis used in current design methods.It is obvious that the load distribution angle, a, increases with increasing basecourse thickness, and based on these data points there is an indication that the variationmight also be attributed to the stiffness of the geosynthetics. It would be reasonable toexpect an upper bound value for a governed by characteristics of the base coursematerial thickness. Most design methods incorporate a single value of load distributionangle, however these results indicate the behavior is more complex.6.6 STRESSES AND BEARING CAPACITY FACTORSMost design methods recognize that, by employing a geosynthetic in unpavedroad construction over soft ground, the subgrade bearing capacity failure is movedfrom being one of local shear failure to general shear failure, which is an increase inbearing capacity of approximately 65%.Estimation of vertical stresses can be calculated by several methods. Designmethods usually do not consider the effects of multi-layer system and how the elasticmodules ratio between the layers effect the stress distribution with depth. Haliburtonand Barron (1983) did however realize this effect and concluded that by usingBurmister theory (1943, 1944) of stresses and displacements in layered systems tocalculate the stresses on the subgrade surface, the stresses were half the stresses thatwould be estimated by employing the conventional Boussinesq theory. They howeverChapter Six -Interpretation of Field Data^ 125did not mention that this stress reduction based on the Burmister layered system isfairly sensitive to the modulus ratio between the layers.To demonstrate this stress reduction effect for a two-layer system a simplermethod developed by Odemark (1949) is used, which is based on Burmister's layeredsystem theory. Two different elastic modulus ratios are shown in Fig. 6.16 to illustratethe effects of the two-layered system on stress distribution in the upper layer using astandard axle load and a tire inflation pressure of 620 kPa.111^11111111111111BoussinesqOdemark - El / E2 = 10Odemark - El / E2 = 100Giroud and Noiray11111111111111111111111111111111111111111111111110.0^0.1^0.2^0.3^0.4^0.5^0.6^0.7^0.8^0.9^1.0Base Course Thickness (m)Figure 6.16 Vertical stresses on subgrade surface versus base course thicknessFor comparison, the conventional Boussinesq is also shown, and the pyramidal stressreduction used in Giroud and Noiray's design procedure. It is obvious that there issignificant stress reduction that takes place when two-layer system is considered, andthe 50% stress reduction proposed by Haliburton and Barron (1983) seems to be infairly good agreement with the Odemark theory, though it is dependent on the elasticmodulus ratio. Vertical stresses predicted using a pyramidal stress distribution over-0Chapter Six - Interpretation of Field Data^ 126predict the stresses on subgrade surface compared to the Odemark theory by 25 - 90%,depending on the base course thickness, using a modulus ratio of 10 which is thepredicted ratio in the field trial.The comparison of the unreinforced data in Fig. 6.13 indicated a good agreementat rut depths of 15 cm between the field trial data and the semi-theoretical procedureby Giroud and Noiray. The reinforced comparison indicates an over-estimation inperformance which is difficult to explain assuming that the vertical stresses areoverestimated. Whether this can be attributed to how the tensioned membrane effect istreated, how traffic volume is taken into account in the reinforced system, the fact thatthe base course material has other material properties than CBR of 11, or simply due tothe reason that the ultimate bearing capacity is over estimated, is hard to tell but thereis an indication that some other factors have to be considered.The relationship between rut depth and the predicted vertical stress on thesubgrade surface normalized with respect to the undrained shear strength is shown inFig. 6.17 and Fig 6.18, for the unreinforced section and the reinforced sectionsrespectively. The analysis was made to estimate bearing capacity factors mobilized bythe unreinforced and the reinforced sections. Rut depths are plotted at certain intervalof vehicle passes, and the vertical stresses calculated using the Odemark (1949)method knowing the base course thickness change for various number of passes. Inthe unreinforced case the bearing capacity factor, given by the normalized verticalstress, starts to increase at a value of o-z/cu = 2.2, which is about 70% of the bearingcapacity value used in most design methods in unreinforced systems and is thereforeconsidered to be in a fairly good agreement. The reinforced system was treated in thesame manner as the unreinforced and the result are shown in Fig. 6.18. Unfortunatelythe changes in base course thickness for the 25 cm thick sections is only available up to60 - 80 passes and therefore vertical stresses for passes over that number are excluded18 —16 —14E1210200N=160• N=300N=10N=20^ N=40• N=800■Reinforced —•0■■♦6420 I^I^I^IChapter Six - Interpretation of Field Data^ 12720 I^I^I^I^I^I^I^I^I^0E•r1816141210860o 0 0 00^^0Unreinforced —4• N=10^ N=20O N .4000^1^2^3^4^5^6^7^8^9^10Ratio of Subgrade Stress over Undrained Shear Strength, tAFigure 6.17 Bearing capacity factor prediction - Unreinforced Sections0^1^2^3^4^5^6^7^8^9^10Ratio of Subgrade Stress over Undrained Shear Strength, P-Figure 6.18 Bearing capacity factor prediction - Reinforced sections2Chapter Six - Interpretation of Field Data^ 128and the highest stress level for that range of passes is between 4 and 4.5 which is stillconsiderably less than 5.14 which is considered as being the stress level for the ultimatebearing capacity of the subgrade using theory of plasticity. The point shown in Fig.6.18, having value of about 5.5 is extrapolated from the 25 cm thick section and istherefore not reliable. Other data points are scattered between about 0.8 and 3.5 andindicate that a ultimate bearing capacity of the subgrade was not mobilized. Therefore,it might indicate that the performance of the thinner sections is in fact attributed tosomething else than bearing capacity factor used in design and rather than reaching theultimate bearing capacity the geosynthetics are torn before that stress level is mobilizedresulting in a local shear failure and therefore much lower stress level than anticipated.If this is true, then outward acting shear stresses and the overall stiffness of thereinforced system are of great importance.Chapter 7Conclusions and Recommendations7.1 CONCLUSIONSThe use of geosynthetics in unpaved road construction is well-recognized.However, the relative improvement of a reinforced over unreinforced system dependson structural considerations such as material properties, geometry and loadingconditions. Few full scale field trials have been performed and documented whichcontrast the relative performance of different geosynthetics under controlled vehicleloading. Design methods differ in some respects as to their interpretation of thereinforcement action. Some basic assumptions made in the design therefore requirevalidation through a well-controlled field trial, and observation of the response of theroad system to trafficking. Some constraints of the field trial must be recognized.These results are based on one strength of subgrade. The vehicle loading ischannalized, which may not always be the case in practice. Compaction of the base129Chapter Seven - Conclusions and Recommendations^ 130course layer exerts an important influence on traffickability, and in this field trial thecompaction prior to vehicle loading was likely to be less than that for regular roadconstruction practice. It should also be recognized that this field trial is limited to 500standard axle vehicle passes, and the relative comparison of performance for both theunreinforced section and the four reinforced sections is based on this number of passes.An unpaved road, comprising a base course layer between 25 and 50 cm. thickover a soft subgrade of undrained shear strength approximately 40 kPa, has been usedto contrast the performance of three non-woven geotextiles and a geogrid. Resultsfrom this full-scale field trial indicate a much improved performance from thereinforced sections than the unreinforced section for unpaved roads over soft ground.However the performance of each of the four different geosynthetics was similar withrespect to failure, where failure is defined as an unacceptable rut depth or aserviceability failure. Compression of the base course layer was observed, which alsomight include some lateral spread of the gravel; the corresponding subgradedeformations were much smaller than rut measurements at the base course surface.The performance of the unreinforced section shows reasonable agreement withthe minimum design thickness proposed by Hammitt(1970) at large rut depths but notat small ruts. No significant variation of subgrade properties was observed along thetest site, therefore the behavior at small rut depths is attributed to early compression ofthe base course layer, and the rut depth is relatively independent of base coursethickness. Thereafter rut development is more dependent on base course thickness,and at this point the dependence of rut on thickness revealed a trend with increasingvehicle passes that compares well with Hammitt. However this does not occur until arut of about 15 cm, which in many cases would be unacceptable and considered closeto a serviceability failure.Chapter Seven - Conclusions and Recommendations^ 131The reinforced data compares quite well with theory (Giroud and Noiray, 1981),considering the trend of rut versus number of vehicle passes, but again the theory overpredicts the performance. This difference is attributed in part to the initial compactionof the base course layer: in theory the base course layer is considered incompressible,and ruts at the base course surface are equal to deformations at the subgrade surface.The difference in performance between theory and the field data, is also attributed tothe nature of the tensioned-membrane effect, which would be greater if the assumptionof same magnitude of ruts and subgrade deformations held true. Although thetensioned-membrane effect is believed to have been significant in the 25 and 30 cmsections, in the thicker sections the membrane effect is greatly over predicted and othermechanisms are more likely to contributing to the better performance of the reinforcedsections.The field performance of the reinforced test sections suggests a particularstabilization mechanism was taking place. In the 25 and 30 cm thick sections thetensioned-membrane effect seems to be mobilized, which is specially evident in the25 cm section where there is a clear difference in performance between the geotextilesthat is related to tensile strength. Separation appears to be very important in thethinnest sections, and the geogrid does not perform so well: as base course thicknessincreases, the geogrid performance gets much better, which is attributed tomaintenance of a good interface bond with the base course material. As base coursethickness increases, the evidence of a tensioned-membrane effect is negligible at thegiven number of vehicle passes and moderate rut depths: stiffness and separation arethe main factors contributing to the better performance of the reinforced system. In the40 and 50 cm thick sections the geogrid consistently shows less rutting than thegeotextiles. While this may be attributed to its ability to take up outward acting shearstresses on the surface of the subgrade below the vehicle wheel paths, all of theChapter Seven - Conclusions and Recommendations^ 132geosynthetics behave in a similar manner. This similar performance might be aconsequence of the geosynthetics exceeding a threshold stiffness for these siteconditions and vehicle loading.7.2 IMPLICATIONS FOR ENGINEERING DESIGNThe currently most widely used design method for unpaved roads over softground is the Giroud and Noiray (1981) procedure. Therefore it is important torecognize the observed over-prediction in performance at large numbers of vehiclepasses, which increases with decreasing traffic volume. This difference is, as alreadystated, largely attributed to the initial base course compaction and leads to the questionof whether or not the base course compaction is realistic in unpaved road design, aswell as the relationship between constant base course thickness and number of vehiclepasses.Based on this field trial, it appears important to consider both stiffness of thegeosynthetic and its separation ability, in design of an unpaved road over a softsubgrade for a small number of vehicle passes and a small base course thickness. It isclear that if proper separation is not provided, the stiffness alone becomes lessimportant, i.e., a geosynthetic that provides a good seperation is a better contributor tothe road system than a stiffer material which does not provide good seperation. Atlarge numbers of passes, and a large base course thickness, separation becomes themost important stabilization factor assuming a threshold stiffness is provided.Based on an equivalent performance, where equivalent performance is defined asthe same rut depth at the same number of vehicle passes, there is potential forconsiderable savings in base course aggregate thickness in a reinforced system, withinRut depth = 15 cmO Unreinforced• Texel Geo 9^ Pohjfelt TS 700▪ Tenser BX 1100• Polyfee TS 6000III^ i^i^I^1^1^1^1^145403530252015 =10 =5 =0 - 00 ■ •■ 0^•Chapter Seven - Conclusions and Recommendations^ 1335045 =4035 =30 =252015 =10 =5 =,4^I^1♦6111^ I^1 111o ■ 0^■o ^ •^■O ^0 ■o ■Rut depth = 5 cm• Unreinforced• Texel Geo 9O Pclyfelt TS 700▪ Tenser BX 1100O PolYfee TS 600•0^ 1 1 I^ \" 11 10 100 500Number of Passes, N504540 =_353025 =201510 =5 =• Unreinforced• Texel Geo 9^ Pdyfelt TS 700▪ Tenser BX 1100• Polyfelt TS 6000^ III1 10^100^5001^I^I^I^I^I^i^tlRut depth = 10 cm 0dh = base coursethickness savings04Number of Passes, N50 I^I^I^I^i^i^l I^lit!1 10 100 500Number of Passes, NFigure 7.1 Base Course Thickness versus Number of Passes - Cost Savings EstimationChapter Seven - Conclusions and Recommendations^ 134an acceptable range of rut depths. Fig. 7.1 shows the base course thickness versusnumber of vehicle passes for three different rut depths, for all five test sections.Savings from use of geosynthetic can be detected from the figure, for the 10 and 15 cmrut depths, whereas behavior at the 5 cm rut depth is dominated by compaction of thebase course and there is no noticeable difference between the unreinforced section andthe reinforced sections. Cost savings are a function of the cost of the geosyntheticsand the base course gravel, and their placement, and for rut depths of 10 and 15 cmthese savings can be estimated from the figure. If however the acceptable rut is lessthan 10 cm, Figs. 6.1 to 6.5 may be used as a guide for cases when a low number ofvehicle passes will be applied, to estimate what kind of geosynthetic would be mosteconomical to use for a given base course thickness. It can be seen that the relativeimprovement is essentially independent of the geosynthetic type for the rut depths of10 and 15 cmIn order to achieve the best performance of a reinforced unpaved road theconstruction has to be carried out differently than for regular unpaved roads. It isimportant that, after the first lift of the base course aggregate placement, compaction iscarried out in such manner that the geosynthetic be placed lightly in tension to bettermobilize the effect of the geosynthetic. Another important factor is to start placementof the gravel and initial compaction from the edges of the road and continue to fill inand compact from the outsides towards the middle. This procedure will result in betteranchorage as well as small tension in the geosynthetics is initiated.BibliographyAas, G., Lacasse, S., Lunne, T., and Hoeg, K. (1986), \"Use of In-Situ Tests forFoundation Design on Clay\", Proceedings of the Conference on Use of In-SituTests in Geotechnical Engineering, ASCE Special Publication 6, pp. 1-30.Alenowicz, J., and Dembicki, E. (1991), \"Recent Laboratory Research on UnpavedRoad Behavior, Geotextiles and Geomembranes, Vol.10, pp. 21-34Armstrong, J.E. (1984), \"Environmental and Engineering Applications of the SurficialGeology of the Fraser Lowland, British Columbia\", Paper 83-23, GeologicalSurvey of Canada.Armstrong, J.E. (1956), \"Surficial Geology of Vancouver Area, British Columbia\",Paper 55-40, Geological Survey of Canada.Armstrong, J.E. (1957), \"Surficial Geology of New Westminster Map-Area, BritishColumbia\", Paper 57-5, Geological Survey of Canada.Austin, D.N., and Coleman, D.M. (1993), \"A Field Evaluation of Geosynthetic-Reinforced Haul Roads Over Soft Foundation Soils\", Conference Proceedings,Geosynthetics '93, Vancouver, Canada.Bakker, J.G. (1977), \"Mechanical Behavior of Membranes in Road Foundations\",Proc. International Conference on the Use of Fabrics in Geotechnics, Paris,Vol.!, pp. 139-142.Barksdale, R., Robnett, Q., and Lai, J. (1982), Proceedings, Second InternationalConference on Geotextiles, Las Vegas, USA, Vol. II, p375.Bauer, A., and Preissner, H. (1986), \"Studies on the Stability of a GeotextileReinforced Two-Layer System\", Proceedings, Third International Conference onGeotextiles, Vienna, Austria, pp. 1079-1084.135Bibliography^ 136Becker, D.E., Crooks, J.H.A., and Been, K. (1988), \"Interpretation of the Field VaneTest in Terms of In-Situ and Yield Stresses\", Vane Shear Strength Testing inSoils:Field and Laboratory Studies, ASTM STP 1014, A.F. Richards, Ed.,American Society for Testing and Materials, Philadelphia, pp. 71-87.Bender, D.A., and Barenberg, E.J.(1978), \"Design and Behavior of Soil-Fabric-Aggregate Systems\", Transportation Research Record 671, TRB, Washington,pp. 64-75.Bjerrum, L. (1973), \"Problems of Soil Mechanics and Construction on Soft Clays\",Proceedings of the 8th International Conference on Soil Mechanics andFoundation Engineering, Vol.3, pp. 111-159.Blunden, R.H. (1973), \"Urban Geology of Richmond, British Columbia\", Departmentof Geology, University of British Columbia, Report No.15.Blunden, R.H. (1975), \"Historical Geology of the Lower Fraser River Valley\",University of British Columbia, Adventures in Earth Science Series, No.3.Bonaparte, R., and Christopher, B.R. (1987), \"Design and Construction of ReinforcedEmbankments Over Weak Foundations\", Transportation Research Record,No.1153, pp. 26-39.Bonaparte, R., and Berg, R. (1987), \"Long-Term Allowable Tension for GeosyntheticReinforcement\", Proceedings of Geosynthetics '87 Conference, New Orleans,Louisiana, USA, pp. 181-192.Bourdeau, P.L., Miskin, K.K., and Fuller, J.M. (1990), \"Behavior of Geotextile-Reinforced Soil Under Cyclic Loading\", Proceedings, 4th InternationalConference on Geotextiles, Geomembranes and Related Products, Hague,Netherlands.Bourdeau, P.L., Holtz, R.D., and Chapuis, J.(1988), \"Effect of Anchorage andModulus in Geotextile-Reinforced Unpaved Roads\", Geotextiles andGeomembranes, Vol.7, pp. 221-230.Bibliography^ 137Bourdeau, P.L., Harr, M.E., and Holtz, R.D. (1982), \"Soil-Fabric Interaction. AnAnalytical Model\", Proceedings, Second International Conference onGeotextiles, Las Vegas, USA, Vol. II, pp. 387-391.Brorsson, I., and Eriksson, L. (1986), \"Long-Term Properties of Geotextiles and TheirFunction as a Separator in Road Construction\", Proceedings, Third InternationalConference on Geotextiles, Vienna, Austria.Burd, H.J., and Brocklehurst, C.J. (1990), \"Finite Element Studies of the Mechanics ofReinforced Unpaved Roads\", Proceedings, 4th International Conference onGeotextiles, Geomembranes and Related Products, Hague, Netherlands, pp. 217-221.Burd, H.J., and Houlsby, G.T. (1989), \"Numerical Modeling of Reinforced UnpavedRoads\", Third Int. Symp. on Numerical Models in Geomechanics, Canada.Burmister, D.M. (1944), \"The General Theory of Stresses and Displacements inLayered Systems\", Journal of Applied Physics, Vol.16, pp. 89-94, 126-127, 296-302.Burmister, D.M. (1943), \"The Theory of Stresses and Displacements in LayeredSystems and Applications to the Design of Airport Runways\", Highway ResearchBoard, Proceedings of the 23rd Annual Meeting, Illinois, pp. 126-144.Candler, R.J. (1988), \"The In-Situ Measurements of the Undrained Shear Strength ofClays Using the Field Vane\", Vane Shear Strength Testing in Soils:Field andLaboratory Studies, ASTM STP 1014, A.F. Richards, Ed., American Society forTesting and Materials, Philadelphia, pp. 13-45.Christopher, B.R., and Holtz, R.D. (1985), \"Geotextile Engineering Manual\",published by the Federal Highway Administration, Report Number FHWA-TS-86/203, Washington, D.C., USA.Christopher, B.R., and Holtz, R.D. (1991), \"Geotextiles for Subgrade Stabilization inPermanent Roads and Highways\", Proceedings of Geosynthetics '91 Conference,Atlanta, USA, pp. 701-713.Bibliography^ 138Christopher, B.R., Holtz, R.D., and Bell, D. (1986), \"New Tests for Determining theIn-Soil Stress-Strain Properties of Geotextiles\", Proceedings, Third InternationalConference on Geotextiles, Vienna, Austria, pp. 683-688.Clague, J.J., and Luternauer, J. (1982), \"Late Quaternary Sedimentary Environments,Southwestern British Columbia\", International Association of Sedimentologists,11th International Congress, Hamilton.Clague, J.J., and Luternauer, J. (1983), \"Late Quaternary Geology of SouthwesternBritish Columbia\", Geological Association of Canada, Field Trip GuidebookNo.10.Craig, R.F. (1992), \"Soil Mechanics\", 5th Edition, Chapman & Hall, New York.Dawson, A.R., and Little, P.H. (1990), \"Reinforced Haul-Roads: Trials at Bothkennar,Scotland\", Proceedings, 4th International Conference on Geotextiles,Geomembranes and Related Products, Hague, Netherlands.De Gardiel, R., and Javor, E. (1986), \"Mechanical Reinforcement of Low-VolumeRoads by Geotextiles\", Proceedings, Third International Conference onGeotextiles, Vienna, Austria, pp. 1021-1026.De Groot, M., Janse, E., Maagdenburg, T.A.C., and Van den Berg, C. (1986), \"DesignMethod and Guidelines for Geotextile Applications in Road Construction\",Proceedings, Third International Conference on Geotextiles, Vienna, Austria,p741.Delmas, Ph., Matichard, Y., Gourc, J.P., and Riondy, G. (1986), \"Unsurfaced RoadsReinforced by Geotextiles-A Seven Years Experiment\", Proceedings, ThirdInternational Conference on Geotextiles, Vienna, Austria, pp. 1015-1020.Dembicki, E., and Jermolowicz, P. (1991), \"Soil-Geotextile Interaction\", Geotextilesand Geomembranes, Vol.10, pp. 249-268Bibliography^ 139Douglas, R.A., Bessey, R.B., and Small, R.P. (1985), \"The Use of Geotextiles inForest Road Construction\", Proceedings, Second Canadian Symposium onGeotextiles and Geomembranes, Edmonton, pp. 89-96.Douglas, R.A., and Kelly, M.A. (1986), \"Geotextile Reinforced Unpaved LoggingRoads: The Effect of Anchorage\", Geotextiles and Geomembranes, Vol.4, pp.93-106.Douglas, R.A., and Valsangkar, A.J. (1992), \"Unpaved Geosynthetic-Built ResourceAccess Roads:Stiffness Rather than Rut Depth as the Key Design Criterion,Geotextiles and Geomembranes, Vol.11, pp. 45-59.Douglas, R.A. (1992), \"Anchorage and Modulus in Geotextile-Reinforced UnpavedRoads\", Geotextiles and Geomembranes, Vol.11, pp. 99-111.Douglas, R.A. (1990), \"Anchorage and Modulus in Geotextile-Reinforced UnpavedRoads\", Geotextiles and Geomembranes, Vol.9, pp. 261-267.Drucker, C.S., and Prager, W. (1952), \"Soil Mechanics and Plastic Analysis or LimitDesign\", Quart. Appl. Math. 10, pp. 157-165.Eisbacher, G.H. (1977), \"Vancouver Geology\", Geological Association of Canada,Vancouver.El-Fermaoui, A., and Nowatzki, E. (1982), \"Effect of Confining Pressure onPerformance of Geotextiles in Soils\", Proceedings, Second InternationalConference on Geotextiles, Las Vegas, USA, Vol. III, pp. 799-804.Floss, R., Laier, H., and Brau, G. (1990), \"Dynamic Loading of Geotextile/Soil-Systems\", Proceedings, 4th International Conference on Geotextiles,Geomembranes and Related Products, Hague, Netherlands, pp. 183-188.Floss, R., and Gold, G. (1990), \"Use of FEM for the Single Reinforced Two-LayerSystem\", Proceedings, 4th International Conference on Geotextiles,Geomembranes and Related Products, Hague, Netherlands.Bibliography^ 140Giroud, J.P., and Noiray, L. (1981), \"Geotextile-Reinforced Unpaved Road Design\",Journal of the Geotechnical Division, ASCE, Vol.107, GT9, Sept., pp. 1233-1254.Giroud, J.P., Ah-Line, C., and Bonaparte, R. (1984), \"Design of Unpaved Roads andTrafficked Areas and Geogrids\", Symposium on Polymer Grid Reinforcement inCivil Engineering, 4.1 London, 23-24 Mars, pp. 1-12.Gourc, J.P., Perrier, H., and Riondy, G. (1983), \"Unsurfaced Road on Soft Subgrade:Mechanism of Geotextile Reinforcement\", Proceedings of the 8th EuropeanConference on Soil Mechanic and Foundation Engineering, Vol.2, Helsinki, pp.495-498.Gourc, J.P., Matichard, Y., Perrier, H., and Delmas, P. (\"1982), Bearing Capacity of aSand Subgrade System with Geotextile\"„ Proceedings, Second InternationalConference on Geotextiles, Las Vegas, USA, Vol. II, p411.Greig, J.W., Campanella, R.G., and Robertson P.K. (1988), \"Comparison of FieldVane Results with Other In-Situ Test Results\", Vane Shear Strength Testing inSoils:Field and Laboratory Studies, ASTM STP 1014, A.F. Richards, Ed.,American Society for Testing and Materials, Philadelphia, pp. 247-263.Gryczmanski, M., and Sekowski, J. (1986), \"A Composite Theory Application forAnalysis of Stresses in a Subsoil Reinforced by Geotextile\", Proceedings, ThirdInternational Conference on Geotextiles, Vienna, Austria, pp. 181-186.Guido, V.A., Bakkala, C.H., and Nocera, J.J. (1991), \"Eccentric Plate Loading Testson Geogrid-Reinforced Subgrades\", Proceedings of Geosynthetics '91Conference, Atlanta, USA, pp. 381-394.Guido, V.A., Biesiadecki, G.L., and Sullivan, M.J. (1985), \"Bearing Capacity of aGeotextile-Reinforced Foundation\", Proceedings of the Eleventh InternationalConference on Soil Mechanics and Foundation Engineering, San Francisco,Vol.3, pp. 1777-1780.Bibliography^ 141Guido, V.A., Knueppel, J.D., and Sweeney, M.A. (1987), \"Plate Loading Tests onGeogrid-Reinforced Earth Slabs\", Proceedings of the Geosynthetics '87Conference, New Orleans, Vol.1, pp. 216-225.Haas, R. (1985), \"Structural Behavior of Tensar Reinforced Pavements and Some FieldApplications\", Symposium on Polymer Grid Reinforcement in Civil Engineering,London, pp. 166-170.Haas, R., Walls, J., and Carroll, R.G. (1989), \"Geogrid Reinforcement of GranularBases in Flexible Pavements\", Transportation Research Record, No. 1188. pp.19-27.Haliburton, T.A., and Barron, J.V. (1983), \"Optimum-Depth Method of Design forFabric-Reinforced Unsurfaced Roads\", Transpn. Res. Rec. No 916,Transportation Research Board, Washington, D.C., pp. 26-32.Haliburton, T.A., and Fowler, J. (1980), \"Design and Construction of FabricReinforced Roads and Embankments on Soft Foundation\", Proc. 1st CanadianSymposium on Geotextiles, Montreal, pp. 138-147.Hammit, G.M. (1970), \"Thickness Requirements for Unsurfaced Roads and Airfields -Bare Base Support\", Technical Report S-70-5, U.S. Army Engineers, WaterwaysExperiment Station, Vicksburg, Miss., July.Harrison W.J., and Gerrard, C.M. (1972), \"Elastic Theory Applied to ReinforcedEarth\", Journal of the Soil Mechanics and Foundations Division, Vol.98,No.SM7, pp. 1325-1345.Hausmann, M.R. (1986), \"Fabric Reinforced Unpaved Road Design Methods-Parametric Studies\", Proceedings, Third International Conference on Geotextiles,Vienna, Austria, pp. 19-23.Hausmann, M.R. (1987), \"Geotextiles for Unpaved Roads-Review of DesignProcedures\", Geotextiles and Geomembranes 5, pp. 201-233.Bibliography^ 142Hill, R. (1963), \"Elastic Properties of Reinforced Solids: Some Theoretical Principles\",Journ. Mech. Phys. Solids, Vol. 11, p357.Hirano, I., Itho, A., Itho, M., Kawahara, S., Shirasawa, M., and Shimizu, H. (1990),\"Test on Traffickability of a Low Embankment on Soft Ground Reinforced withGeotextiles\", Proceedings, 4th International Conference on Geotextiles,Geomembranes and Related Products, Hague, Netherlands, pp. 227-232.Hird, C.C., and Kwok, C.M. (1990), \"Parametric Studies of the Behavior of aReinforced Embankment\", Proceedings, 4th International Conference onGeotextiles, Geomembranes and Related Products, Hague, Netherlands, pp. 137-142.Holtz, R.D., and Sivakugan. (1987), \"Design Charts for Roads with Geotextiles\",Geotextiles and Geomembranes, Vol.5, No.3, pp. 191-199.Houlsby, G.T., and Jewell, R.A. (1990), \"Design of Reinforced Unpaved Roads forSmall Rut Depths\", Proceedings, 4th International Conference on Geotextiles,Geomembranes and Related Products, Hague, Netherlands, pp. 171-176.Houlsby, G.T., and Jewell, R.A. (1988), \"Analysis of Unreinforced and ReinforcedEmbankments on Soft Clays by Plasticity Theory\", Proceedings, 6th InternationalConference on Numerical Methods in Geomechanics, Innsbruck, 2, pp. 1443-1448.Ingold, T.S., and Miller, K.S. (1986), \"Short, Intermediate, and Long Term Stability ofGeotextile Reinforced Embankments Over Soft Clays\", Proceedings, ThirdInternational Conference on Geotextiles, Vienna, Austria, pp. 337-342.Ingold, T.S., and Crowcroft, P. (1984), \"The Notation of Geotextiles as Separators inRoads\", Ground Engineering, Vol.17, No.1, January.Jaecklin, F.P. (1986), \"Design of Road Base and Geotextile by Regression Analysisfrom Experience Data Sources\", Proceedings, Third International Conference onGeotextiles, Vienna, Austria, pp. 985-990.Bibliography^ 143Janet, P.M. (1986), \"Load Tests on Geogrid Reinforced Gravel Fills Constructed onPeat Subgrades\", Proceedings, Third International Conference on Geotextiles,Vienna, Austria, pp. 87-92.Jarrett, P.M. (1985), \"Evaluation of Geogrids for Construction of Roadways OverMuskeg\", Symposium on Polymer Grid Reinforcement in Civil Engineering,London, pp. 149-153.Jarrett, P.M., Lee, R.A., and Ridell, D.V.B. (1977), \"The Use of Fabrics in RoadPavements Constructed on Peat\", Proc. International Conference on the Use ofFabrics in Geotechnics, Paris, Vol. 1, pp. 19-22.Jessberger, H.L. (1977), \"Load-Bearing Behavior of a Gravel Subbase-NonwovenFabric-Soft Subgrade System\", Proc. International Conference on the Use ofFabrics in Geotechnics, Paris, Voll, pp. 9-13.Jewell, R.A. (1982), \"A Limit Equilibrium Design Method for ReinforcedEmbankments on Soft Foundations\", Proceedings, Second InternationalConference on Geotextiles, Las Vegas, USA, Vol.2, pp. 677-682.Jewell, R.A. (1988), \"The Mechanics of Reinforced Embankments on Soft Soils\",Geotextiles and Geomembranes 7, pp. 237-273.Kaniraj, S.R. (1988), \"Design of Reinforced Embankments on Soft Soils\",Geotechnical Engineering, Vol.19, pp. 127-142.Kasarnovsky, V.D., Polunovsky, A.G., and Brantman, B.P. (1982), \"Design of aTemporary Road Structure with the Use of Textile Membrane\", Proceedings,Second International Conference on Geotextiles, Las Vegas, USA, Vol. II, p371.Kern, C.B., and Buchanan, R.G. (1984), \"Annacis Project Richmond, East-WestFreeway - Interim Geotechnical Report\", Ministry of Transportation andHighways, Geotechnical and Materials Branch, Victoria, B.C., Canada.Bibliography^ 144Kinney, T., and Barenberg, E. (1982), \"The Strengthening Effect of Geotextiles onSoil-Geotextile-Aggregate Systems\", Proceedings, Second InternationalConference on Geotextiles, Las Vegas, USA, Vol. II, p347.Koerner, R.M. (1990), \"Designing with Geosynthetics\", 2nd Edition, Prentice HallPubl. Co., Englewood Cliffs, NJ.Kothari, V.K., and Das, A. (1992), \"Compressional Behavior of NonwovenGeotextiles\" Geotextiles and Geomembranes, Vol.11, pp. 235-252.Kysela, Z. (1986), \"Geotextiles as a Structural Member for Load Redistribution\",Proceedings, Third International Conference on Geotextiles, Vienna, Austria, pp.239-244.Laier, H., and Brau, G. (1986), \"The Use of Geotextiles in Road Construction UnderIntensive Dynamic Loading\", Proceedings, Third International Conference onGeotextiles, Vienna, Austria, pp. 995-1000.Landva, A.O. (1980), \"Vane Testing in Peat\", Canadian Geotechnical Journal, Vol.17,No.1, National Research Council of Canada.Lefebvre, G., Ladd, C.C., and Pare, J.J. (1988), \"Comparison of Field Vane andLaboratory Undrained Shear Strength in Soft Sensitive Clays\", Vane ShearStrength Testing in Soils:Field and Laboratory Studies, ASTM STP 1014, A.F.Richards, Ed., American Society for Testing and Materials, Philadelphia, pp.233-246.Ling, H.I., Wu, J.T.H., and Tatsuoka, F. (1991), \"Effectiveness of In-Membrane Testin Simulating Strength and Deformation Characteristics of a NonwovenGeotextile Under Operational Conditions\", Proceedings of Geosynthetics '91Conference, Atlanta, USA, pp. 601- 614.Love, J.T., Burd, H.J., Milligan, G.W.E., and Houlsby, G.T. (1987), \"Analytical andModel Studies of Reinforcement of a Granular Layer on a Soft Clay Subgrade\",Canadian Geotechnical Journal, Vol.24, No.4, pp. 611-622.Bibliography^ 145Mahmoud, F.F., and Mashhour, M.M. (1986), \"An Elastoplastic Finite Element for theAnalysis of Soil-Geotextile Systems\", Proceedings, Third InternationalConference on Geotextiles, Vienna, Austria, pp. 229-231.McGown, A., Andrawes, K.Z., and Kabir, M.H. (1982), \"Load-Extension Testing ofGeotextiles Confined In-Soil\", Proceedings, Second International Conference onGeotextiles, Las Vegas, USA, Vol. III, pp. 793-798.McGown, A., Andrawes, K.Z., and Yeo, K.C. (1985), \"The Load-Strain-TimeBehaviour of Tensar Geogrids\", Symposium on Polymer Grid Reinforcement inCivil Engineering, London, pp. 11-17.Meyerhof, G.G. (1974), \"Ultimate Bearing Capacity Of Footings on Sand LayerOverlaying Clay\", Canadian Geotechnical Journal, Vol.11, No.2, pp. 223-229.Milligan, G.W.E., Jewell, R.A., Houlsby, G.T., and Burd, H.J. (1989), \"A NewApproach to the Design of Unpaved Roads-Part I\", Ground Engineering, Vol.22,No.3, pp. 25-29.Milligan, G.W.E., Fannin, J., and Farrar, D.M. (1986), \"Model and Full-Scale Tests ofGranular Layers Reinforced with a Geogrid\", Proceedings, Third InternationalConference on Geotextiles, Vienna, Austria, pp. 61-66.Milligan, G.W.E., and Love, J.P. (1984), \"Model Testing of Geogrids Under anAggregate Layer on Soft Ground\", Symp. on Polymer Grid Reinforcement inCivil Engineering, Paper No 4.2, London, (ICE).Milligan, G.W.E., Jewell, R.A., Houlsby, G.T., and Burd, H.J. (1989), \"A NewApproach to the Design of Unpaved Roads-Part II\", Ground Engineering,Vol.22, No.8, pp. 37-42.Milligan, V., and Rochelle, P.La. (1985), \"Design Methods for Embankments OverWeak Soils\", Symposium on Polymer Grid Reinforcement in Civil Engineering,London, pp. 95-102.Bibliography^ 146Milligan, G.W.E., and Love, J.P. (1985) \"Model Testing of Geogrids Under anAggregate Layer on Soft Ground\", Symposium on Polymer Grid Reinforcementin Civil Engineering, London, pp. 128-138.Miura, N., Sakai, A., and Taesiri, Y. (1990), \"Polymer Grid Reinforced Pavement onSoft Clay Grounds\", Geotextiles and Geomembranes, Vol.9, pp. 99-123Murray, R.T., and McGown, A. (1990), \"Geotextiles in Road Construction: Summaryof an OECD Report\", Proceedings, 4th International Conference on Geotextiles,Geomembranes and Related Products, Hague, Netherlands.Mylleville, B.L.J., and Rowe, R.K. (1991), \"On the Design of ReinforcedEmbankments on Soft Brittle Clays\", Proceedings of Geosynthetics '91Conference, Atlanta, USA, pp. 395-408.Nieuvenhuis, J.D. (1977), \"Membranes and the Bearing Capacity of Road Bases\",Proc. International Conference on the Use of Fabrics in Geotechnics, Paris,Vol.1, pp. 3-8.Nishigata, T., and Yamaoka, I. (1990), \"Bearing Capacity of Unpaved RoadReinforced by Geotextile\", Proceedings, 4th International Conference onGeotextiles, Geomembranes and Related Products, Hague, Netherlands.Odemark, N. (1947), \"Undersäkning av elasticitetsegenskaperna hos olika jordartersaint teori for berakning av belaggningar enligt elasticitetsteorin\", StatensVaginstitut, Stockholm, Sweden.Petrik, P.M. (1977), \"Development of Stresses in Reinforcement and Subgrade of aReinforced Soil Slab\", Proc. International Conference on the Use of Fabrics inGeotechnics, Paris, Vol.1, pp. 151-154.Potter, J.F., and Currer, E.W.H. (1981), \"The Effect of a Fabric Membrane on theStructural Behavior of a Granular Road Pavement\", Transport and RoadResearch Laboratory, TRRL Laboratory Report 996.Bibliography^ 147Purushothamaraj, P., Ramiah, B.K., and Venkatakrishna Rao, K.N. (1973), \"BearingCapacity of Strip Footings in Two Layered Cohesive-friction Soils\", CanadianGeotechnical Journal, Vol.11, No.32, 1974, pp. 32-45.Ramalho-Ortiago, J.A., and Palmeia, E.M. (1982), \"Geotextile Performance at anAccess Road on Soft Ground near Rio de Janeiro\", Proceedings, SecondInternational Conference on Geotextiles, Las Vegas, USA, Vol. II, p353.Raumann, G. (1982), \"Geotextiles in Unpaved Roads: Design Considerations\",Proceedings, Second International Conference on Geotextiles, Las Vegas, USA,Vol. II, pp. 417-422.Resl, S., and Werner, G. (1986), \"The Influence of Nonwoven Needle punchedGeotextiles on the Ultimate Bearing Capacity of the Subgrade\", Proceedings,Third International Conference on Geotextiles, Vienna, Austria, pp. 1009-1013.Robnett, Q., Lai, J., and Murch, L. (1982), \"Effect of Fabric Properties on thePerformance and Design of Aggregate-Fabric-Soil Systems\", Proceedings,Second International Conference on Geotextiles, Las Vegas, USA, Vol. II, p381.Robnett, Q., and Lai, J.S. (1982), \"Fabric-Reinforced Aggregate Roads-Overview\",Transportation Research Record, No. 875. pp. 42-50.Robnett, Q.L., Lai, J.S., Lavin, J.G., Murch, L.E., and Murrey, C.D. (1980), \"Use ofGeotextiles in Road Construction:An Experimental Study, Proc. 1st CanadianSymposium on Geotextiles, Montreal, pp. 113-124.Roy, M., and Leblanc, A. (1988), \"Factors Affecting the Measurements andInterpretation of the Vane Strength in Soft Sensitive Clays\", Vane ShearStrength Testing in Soils:Field and Laboratory Studies, ASTM STP 1014, A.F.Richards, Ed., American Society for Testing and Materials, Philadelphia, pp.117-128.Ruddock. E.C., Potter, J.F., and McAvoy, A.R. (1982), \"A Full-Scale Experiment onGranular and Bituminous Road Pavements Laid on Fabrics\", Proceedings,Second International Conference on Geotextiles, Las Vegas, USA, Vol. II, p365.Bibliography^ 148Sawicki, A. (1983), \"Plastic Limit Behavior of Reinforced Earth\", Journal of theGeotechnical Engineering Division, ASCE, Vol.107, No 7, p1000.Sellmeijer, J.B., Kenter, C.J., and Van den Berg, C. (1982), \"Calculation Method for aFabric Reinforced Road\", Proc. SICG, 2, Aug., pp. 393-398.Sellmeijer, J.B. (1990), \"Design of Geotextile Reinforced Paved Roads and ParkingAreas\", Proceedings, 4th International Conference on Geotextiles,Geomembranes and Related Products, Hague, Netherlands.Sellmeijer, J.B., Kenter, C.J., and Van den Berg, C. (1982), \"Calculation Method for aFabric Reinforced Road\", Proceedings, Second International Conference onGeotextiles, Las Vegas, USA, Vol. II.Silvestri, V., and Auberni, M. (1988), \"Anisotropy and In-Situ Vane Tests\", VaneShear Strength Testing in Soils:Field and Laboratory Studies, ASTM STP 1014,A.F. Richards, Ed., American Society for Testing and Materials, Philadelphia, pp.89-103.Sorlie, A. (1977), \"The Effects of Fabrics on Pavement Strength-Plate bearing Tests inthe Laboratory\", Proc. International Conference on the Use of Fabrics inGeotechnics, Paris, Vol.1, pp. 15-18.Sowers, G.F., et al. (1982), \"Mechanism of Geotextile-Aggregate Support in Low-Cost Roads\", Proceedings, Second International Conference on Geotextiles, LasVegas, USA, Vol. II, p341.Steward, J., Williamson, R., and Mohney, J. (1978), \"Guidelines for Use of Fabrics inConstruction and Maintenance of Low-Volume Roads\", USDA, Forest Service,Portland, Oregon. Also published as Report No. FHA-TS-78-205.Steward, J., and Mohney, J. (1982), \"Trial Use Results and Experience UsingGeotextiles for Low-Volume Forest Roads\", Proceedings, Second InternationalConference on Geotextiles, Las Vegas, USA, Vol. II, p335.Bibliography^ 149Steward, J.E., Williamson, R., and Mohney, J. (1983), \"Guidelines for Use of Fabricsin Construction and Maintenance of Low-Volume Roads\", Federal HighwayAdministration, Report Number FHWA-TS-78-205.Terzaghi, K. (1943), \"Theoretical Soil Mechanics\", John Wiley and Sons, Inc., NewYork.Terzaghi, K., and Peck, R.B. (1967), \"Soil Mechanics in Engineering Practice\", 2ndEdition, John Wiley and Sons, Inc., New York.Van Zanten., Ed. (1986), \"Geotextiles and Geomembranes in Civil Engineering\",Wiley-Halsted.Watari, Y., Higuchi, Y., and Aboshi, H. (1986), \"Field Studies of the Behavior ofGeogrids and Very Soft Ground\", Proceedings, Third International Conferenceon Geotextiles, Vienna, Austria, pp. 187-191.Webster, S.L., and Watkins, J.E. (1978), \"Investigation of Construction Concepts forPavements Across Soft Ground\", Technical Report S-78-6, U.S. ArmyEngineers, Waterways Experiment Station, Vicksburg, Miss., July.Webster, S.L., and Watkins, J.E. (1977), \"Investigation of Construction Techniques forTactical Bridge Approach Roads Across Soft Ground\", Technical Report S-77-1,U.S. Army Engineers, Waterways Experiment Station, Vicksburg, Miss.,February.Wroth, C.P. (1984), Geotechnique, Vol.34, pp. 449-489."@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "1993-11"@en ; edm:isShownAt "10.14288/1.0050484"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Civil Engineering"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Geosynthetic stabilization of unpaved roads on soft ground : a field evaluation"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/1462"@en .