ARCHIVESOFCIVILENGINEERING,LVIII,1,2012 DOI:10.2478/v.10169-012-0006-z THE INFLUENCE OF THE CARBO-GLASS GEOGRID-REINFORCEMENT ON THE FATIGUE LIFE OF THE ASPHALT PAVEMENT STRUCTURE J. GÓRSZCZYK1, S. GACA2 Thispaperdescribestheanalysesofthefatiguelifeoftheasphaltpavementreinforcedwithgeogrid interlayerundertrafficloading.FiniteElementANSYSpackagewithusingnCodeapplications,as well as macros specially designed in APDL programming script and VBA were used to model the considered problem. Our analysis included computation of stress, fatigue life, damage matrix and rainflow matrix. The method applied was the one of fatigue calculation: stress – number of cyclesinshortS-N.Onthebasisoftheperformedhighcyclefatigueanalysis,theinfluenceofthe locationoftheusedgeogridandofitsbondwithasphaltlayersonthefatiguelifeandtheworkof the asphalt pavement structure were determined. The study was carried out for three temperature seasonsi.e.springandfall(assumedasoneseason),winterandsummer.Thevariabilityofthetraffic conditionsweretakenintoaccountbyassumingweeklyblocksoftrafficloading.Thecalculations were made using the real values of loading measured in field tests on the German highways by means of HS-WIM weighing system. As a result of the performed tests, it was proved that the use of geogrid-reinforcement may prolong the fatigue life of the asphalt pavement. However, it is requiredthat:thegeogridshouldbelocatedinthetensionzoneaslowaspossibleinthestructure of the asphalt layers. Moreover, it is necessary to provide high stiffness of the bond between the geogridandtheasphaltlayers. Keywords: geogrid,asphaltpavements,discreteandfiniteelementmodeling,fatigue,geogrid-reinforcement, contactoflayers,stress–numberofcycles. 1. I A significant growth in heavy load traffic is a currently observed in many European countries. For instance in Poland during the last decade the number of heavy vehicles travelling on the roads has doubled and the permissible single axle load value was also increased. The main European road network has been designed for the axle loading of 115 kN, which for our country is a major challenge, because so far the majority of roads in Poland have been designed for loading of 80 kN and 100 kN per axle. 1 Ph.D. (Corresponding Author), Cracow University of Technology, Kraków, Poland, e-mail: [email protected] 2 Prof.,Ph.D.,CracowUniversityofTechnology,Kraków,Poland,e-mail:[email protected] 98 J.G´,S.G Searching for new ways to increase the durability of their asphalt pavement one should consider their reinforcement with selected kinds of geogrids or composites (a combination of geogrid and unwoven material) which are characteristic for their stiffness and the stiffness of the bonding to the asphalt layers [1], [2], [3], [6], [7], [11], [18]. This feature is a new approach, because the geogrid was almost always applied as the stress relieving interlayer to repair the reflection cracks created as a result of the discontinuities in the lower layers of the pavement structure. Reinforcing the asphalt pavement in such a way that the discontinuity in the lower layers is eliminated results, first of all, in improvement of the fatigue life of the whole structureandsointheincreaseofthenumberoftheloadingscarriedbytheconstruction before it is damaged, or the decrease in the thickness of the designed layer while maintaining the level of desired fatigue life. Since this technology is still in the process of development, there is a number of issues that still wait for consideration. One of them is the question of the choice of the type of geogrid-reinforcement as well as its location, and the method of how to build it in. So far the studies dealing with these issues have had an experimental character and to in less range, the theoretical one [8], [9], [14], [17]. These conditions were taken into account, in certain research on the fatigue life of the asphalt pavement reinforced with a geogrid consisting of the carbo-glass fibre with relation to various kinds of interlayer bonding submitted to traffic loading and with relation to the temperature seasons characteristic for the selected, south region of Poland. The studies correspond to the papers described in the FHWA report [12] on nu- merical analyses of the asphalt pavement reinforced with the geogrid located in the aggregate subbase. However, location of the geogrid in the asphalt layers and the eva- luationofthefatiguelifeofsuchastructurerequiredapplicationofdifferentnumerical methods. 2. N  2.1. V Thesubjectoftheanalysiswastheasphaltpavementstructureforheavytrafficloadi.e. with the daily traffic load of 2000 equivalent axles of 100 kN. The pavement structure was reinforced with a geogrid spread over the entire width of the carriageway. In the numerical model the reinforcement layer was located in three different places i.e. on theaggregatesubbasewiththeuseofthe3cmthickasphaltlevellinglayer(i.e.located at a depth of 31 cm), in the middle of the asphalt base layer (located 22 cm deep) and at the bottom of the asphalt binder layer (at a depth of 13 cm). The asphalt pavement structure presenting the location of the geogrid is shown in Fig. 1. TIC-GG-RFL... 99 Fig.1.Thesystemofthelayersintheanalyzedasphaltpavementshowingthevariantlocationsofthe geogrid-reinforcement. Rys.1.Układwarstwanalizowanejnawierzchnizzaznaczonymiwariantowymilokalizacjamipośredniej warstwyzbrojącej,tj.geosiatki The analyses were carried out for several variations of the pavement structure which differed not only in the location of the geogrid-reinforcing but also in the way of bonding between layers. The variants of the contact conditions between layers used in the models of pavement structure are listed in Table 1. Until now laboratory tests have indicated the decrease in bonding between layers due to the presence of the geogrid in comparison to the bonding without geogrid (what was compensated by growth of fatigue life) [9]. For this reason the contact zones between geogrid and asphalt layers were described by Coulomb’s law. Several values of the friction coefficient were assumed and the fatigue life of the pavement structure in relation to this parameter was determined. The material parameters of the structure layers and subgrade used for the numerical calculations are listed in Table 3. Material and geometric parameters for the geogrid-reinforcement were determined in thelaboratorytestsusinguniaxialtensionmethodofthewidesamplesandextractionof the fibers [5]. The numerical analyses were made for all seven variants of the bonding between layers in three locations of the geogrid in the pavement structure. The values of the numerical analyses are listed below. For comparison purposes additional analyses for the following two cases of the structure without geogrid-reinforcement were carried out. The results are presented in Table 2. 100 J.G´,S.G Table1 Variantsofbondingbetweenlayersoftheasphaltpavementstructurereinforcedwiththe geogrid-reinforcement. Wariantypołączeniamiędzywarstwowegowmodelachkonstrukcjinawierzchnizbrojonejgeosyntetykiem Symbol Typeofbondingbetweenthelayersofthestructure Fullbondingbetweenallthelayersandbetweentheasphaltlayersandthegeo- B–case1 grid-reinforcement(continuityofallthecomponentsofthedisplacementvector) Nobondingbetweenthegeogridandtheasphaltlayersinthetangent direction(bondingonlyinthedirectionnormaltothecontactarea),thecoeffi- F=0.00–case2 cientoftheCoulombmodelfrictionisF=0.00.Theremaininglayersofthe pavementstructurewereinfullbonding. Nobondingbetweenthegeogridandtheasphaltlayersinthetangent direction(bondingonlyinthedirectionnormaltothecontactarea),thecoeffi- F=0.25–case3 cientoftheCoulombmodelfrictionisF=0.25.Theremaininglayersofthe pavementstructurewereinfullbonding. Nobondingbetweenthegeogridandtheasphaltlayersinthetangent direction(bondingonlyinthedirectionnormaltothecontactarea),thecoeffi- F=0.50–case4 cientoftheCoulombmodelfrictionisF=0.50.Theremaininglayersofthe pavementstructure-fullbonding. Nobondingbetweenthegeogridandtheasphaltlayersinthetangent direction(bondingonlyinthedirectionnormaltothecontactarea),thecoeffi- F=0.75–case5 cientoftheCoulombmodelfrictionisF=0.75.Theremaininglayersofthe pavementstructurewereinfullbonding. Nobondingbetweenthegeogridandtheasphaltlayersinthetangent direction(bondingonlyinthedirectionnormaltothecontactarea),thecoeffi- F=1.20–case6 cientoftheCoulombmodelfrictionF=1.20.Theremaininglayersofthe pavementstructurewereinfullbonding. Fullbondingbetweenthegeogridandasphaltlevellinglayer,between thegeogridandthebaselayer-nobondinginthetangentdirection(bonding F=0.00*–case7 onlyinthedirectionnormaltothecontactarea),thecoefficientoftheCoulomb modelfrictionisF=1.20.Theremaininglayersofthestructurewereinfullbon- ding. Table2 Variantsofbondingbetweenlayersoftheasphaltpavementstructurewithoutthegeogrid-reinforcement. Wariantypołączeniamiędzywarstwowegowmodelachkonstrukcjinawierzchniniezbrojonej geosyntetykiem Symbol Typeofbondingbetweenthelayersofthestructure Fullbondingbetweenalllayersofthepavementstructure(continuityofallthecom- ZW–case1 ponentsofthedisplacementvector) Nobondingbetweenasphaltlayers(thefrictioncoefficient =0.00)butwith BZW–case2 thefullbondingbetweenthecontactareaoftheaggregatesubbaseandsubgradeand alsoatthefullbondingbetweenthebaseandaggregatesubbase. 2.2. M   The elastic multi-layers system where the geogrid is modeled as reinforcement by a membranewithonlytensilestiffnesswasassumedasaphysicalmodelofthepavement and subgrade. Besides, it was also assumed that with one exception the full bonding TIC-GG-RFL... 101 does not exist between the membrane and the asphalt layers. The Coulomb friction model in the tangent directions to the surface of the contact was introduced. It is characterized by friction coefficient µ defined as (2.1): τ k (2.1) µ = p k where: τ is shear stress on the contact surface, p is normal stress to the contact k k surface. At the normal direction to the contact surface, the continuity of the components of the displacement vector was assumed. For the remaining layers between which the geogrid was not placed, the full bonding was assumed (the continuity of the displa- cement). The conditions of the geogrid-reinforcement binding with asphalt layers for top and for the bottom of the geogrid were described by one value of the friction coefficient µ. The constitutive models for all materials of which the pavement structure was composed, were assumed for the fatigue analyses as linear elastic (Hooke’s model). The values of the stiffness modulus for the asphalt layers, which are the function of the speed at which the vehicles travel (relating to the time of loading) and the temperature, were assumed on the basis of the temperature profiles and the average speed of the vehicles for specific temperature seasons (Fig. 2). The stiffness modulus of asphalt layers replaces Hooke’s modulus, what is the certain simplification in ana- lyses. Numerical calculations were made for material parameters listed in Table 3. The geogrid-reinforcement was modeled by the equivalent orthotropic continuum layer – a membrane with only tensile stiffness and the thickness of 0.06 mm. Elastic modulus determined in laboratory tests were: E = 150811 MPa (carbon fibers), E = 59092 XX ZZ MPa (glass fibers) [5], where X and Z are the directions of the orthotropic location of glass and carbon fibers consistent with the global coordinates XYZ of the FEM model. Equal values of shear modulus and the major Poisson’s ratio were assumed: G = 0.20 GPa and ν = 0.20 respectively. XZ xz Table3 Materialparametersforthepavementlayersinthespecificseasons: SPRING/FALL,SUMMER,WINTER. Przyjętestałemateriałowewarstwnawierzchnidlaposzczególnychsezonów: WIOSNA/JESIEŃ,LATO,ZIMA StiffnessModulusE Poisson’sratio TheThicknessof Thelayer [MPa] ν theLayer[cm] WearingcourseAC dependantonthetemperature 0.30 5 BinderlayerAC (changingwiththedepth), 0.30 8 Figure2 BaseAC 0.30 18 AggregateSubbase 400 0.35 20 Subgrade 100 0.35 450 102 J.G´,S.G Besides, it was assumed, that in the subgrade at the depth of 450 cm the bedrock was situated and it did not affect the results of analyses. Fig.2.ThegraphicrepresentationoftheelasticmodulusEoftheasphaltlayers,aggregatesubbaseand subgradeforspecifiedtemperatureseasons,distributionsassumedforthenumericalanalyses. Rys.2.WykresywartościmodułówsztywnościsprężystejEwarstwasfaltowych,kruszywaipodłoża gruntowegodlaposzczególnychsezonówtemperaturowych,rozkładyprzyjętedoanaliznumerycznych Forsimplification,thewheelloadswerereducedtoverticalloadonlyneglectingthe horizontal forces, and they were modeled as the pressure of time dependent value and applied to the surface on a disc with the diameter of 0.32 m (Fig. 5a). The variability of the load in specified seasons was closed in blocks and ascribed to three temperature seasons:SPRING/FALL,WINTER,SUMMER.Typesofvehiclesforspecifiedseasons werecollectedinfieldtestsonGermanhighways[15]andpreparedforfatigueanalysis by ICE–flow system (Fig. 3). Types of vehicles in the selected weeks for specified seasons are presented in Fig. 4. Because of the symmetry of the static model of the construction in the fatigue analysis, only a quarter of the three-dimensional model of the pavement structure was taken into consideration, Fig. 5. It was assumed that evaluation of the fatigue life of the asphalt layers structure (N ) consists in determining the smallest number of cycles after which the damage str of at least one of the materials may occur: asphalt concrete (N ), the material of the p geogrid-reinforcement (N ), or the adhesion layer – the bonding between the geogrid s and the asphalt layers(N ). The geogrid stops playing the role of the reinforcement adh TIC-GG-RFL... 103 Fig.3.AnexampleofasampleloadhistoryintheSUMMERtemperatureseason[Pa]forvehiclewheels a)measuredinfieldtests[15],b)preparedforFEM-fatigueanalysisusingtheICE-flowapplication. Rys.3.PrzykładowyfragmentprzebieguczasowegoobciążeniazsezonutemperaturowegoLATO[Pa] odkółporuszającychsiępojazdówa)pomierzonywbadaniachpolowych[15],b)przygotowanydla analizzmęczeniowychMESzwykorzystaniemaplikacjiICE-flow Fig.4.Typesofvehiclesforthespecifiedtemperatureseasonsdeterminedonthebasis ofthefieldtests[15] Rys.4.Strukturarodzajowapojazdówciężkichdlaposzczególnychsezonówtemperaturowych, wyznaczonanapodstawiebadańpolowych[15] 104 J.G´,S.G Fig.5.a)Geometricmodeloftheasphaltpavementstructureshowingtheareawherethecontact pressurefromtheheavyvehicle’stypeisapplied,b)geometricmodelofthepavementstructureand subgradetakenforFEManalyses–theXYZtheglobalcoordinatesystem. Rys.5.a)Modelgeometrycznykonstrukcjinawierzchnizzaznaczonąpowierzchniądziałaniaciśnienia kontaktowegoodoponypojazduciężkiego,b)modelgeometrycznykonstrukcjinawierzchniipodłoża gruntowegoprzyjętydoobliczeń,XYZ–globalnyukładwspółrzędnych in the structure. Therefore the fatigue life of the asphalt structure reinforced with a geogrid can be described by the following equation (2.2): (cid:16) (cid:17) (2.2) N = min N ,N ,N str p s adh At this phase of the study, it was assumed that only the fatigue material properties of the asphalt concrete (N – the minimal value in the set) play the role in evaluation of p the fatigue life of the pavement structure. It was assumed that the conditions of the contact between all the layers of the structure do not change during the service. Also, it was assumed that the fatigue life of the structure does not indicate the period of time until the entire structure is damaged but only the period of time till local fatigue cracks are initiated and macro-cracks appear (the criterion of fatigue damage). The fatigue life of all the variants (N ) was calculated on the basis of the str Palmgren-Miner hypothesis of the linear summing of the fatigue damage for the specified temperature seasons, according to the equations (2.3–2.6): (2.3) D = 2·D +D +D c SPRING SUMMER WINTER FALL TIC-GG-RFL... 105 n 1 (2.4) D = SPRING FALL NT1 n 2 (2.5) D = SUMMER N T2 n 2 (2.6) D = ZIMA N T2 where (2.3–2.6): D is fatigue damage for all temperature seasons [–], D is fatigue c SPRING/FALL damage for temp. season SPRING/FALL [–], D – fatigue damage for temp. SUMMER season SUMMER [–], D is fatigue damage for temp. season WINTR [–], n is WINTER i the number of the load blocks in a given temperature season – the number of weeks [–], N is the damaging number of load blocks (weeks) for a given temp. season [–]. Ti If the sum D in equation (2.3) reaches the value 1, the pavement structure is c damaged according to the assumptions given above. 2.3. TFEM -  The following elements at the ANSYS code were used from the software ANSYS in FE modeling of the pavement structure: 20-node, solid element SOLID186, 8-node SURF154, the contact surfaces between the layers were modeled with 8-node elements of the TARGE170 and CONTA174. The final mesh of the model is presented in Fig. 6a. The geogrid was modeled by the 4-node anizonomic 2-dimentional elements SHELL41, which enabled functioning of this layer only with the membrane stiffness and with the tensile stiffness Fig. 6b. FEM model of the asphalt pavement structure was subjected to empirical veri- fication. The verification was carried out in the Federal Highway Research Institute in Germany for the variant of the structure without the geogrid in the asphalt layers and submitted to the static load of the vehicle [5]. The results were satisfactory, what enabled the authors to perform further analysis employing the FEM model. 2.4. T    The stress-life (SN) relations for the asphalt layers i.e. the wearing course layer, binder layer, base layer were determined on the basis of the fatigue test results for uniaxial tension at controlled force and carried out at the Braunschweig University by Mollen- hauer [10]. These tests were the continuation of the study by Rubach in the 1990’s, [13]. 106 J.G´,S.G Fig.6.ThemeshoftheFEmodelofpavementstructureandsubgradewiththeSOLID186element, b)ThefinalmeshoftheFEmodelofthegeogrid-reinforcementbuiltintotheasphaltpavementwiththe SHELL41element. Rys.6.a)SiatkaMESprzyjętejkonstrukcjinawierzchnidrogowejipodłożagruntowegozgłównym elementemSOLID186,b)OstatecznasiatkaMESgeosyntetycznejwarstwyzbrojącejwbudowanej wukładwarstwnawierzchnizelementemSHELL41 The experimental fatigue data described by the general correlation (2.7) were re- formulated to formula (2.8) by means of Goodman correction (mean stress effects) taking into account the average temperature of the specified temperature profiles: SPRING/FALL, SUMMER, WINTER for the specific layers of the structure. The temperature correction was made according to the algorithm described in the literature [10]. Due to the lack of the following correlation σ = f(σ ), and taking into acco- a m unt the mean stress, the Goodman theory recommended for the situation where the influence of σ on the fatigue life of the material is not established in the laboratory a tests (2.9). (2.7) N = K ·∆σK2 macro 1 (2.8) σa = SF ·K1(cid:48) ·(Nmacro)K2(cid:48)