SZ. FISCHER, F. HORVÁT Szabolcs FISCHER Ferenc HORVÁT … · 22 2011 slovak university of technology sz. fischer, f. horvÁt investigations of the reinforcement and stabilisation

2011 SLOVAK UNIVERSITY OF TECHNOLOGY22

SZ. FISCHER, F. HORVÁT

INVESTIGATIONS OF THE REINFORCEMENT AND STABILISATION EFFECT OF GEOGRID LAYERS UNDER RAILWAY BALLAST

KEY WORDS

• ballast material,• track deterioration,• geogrid reinforcement,• stabilization of railway track geometry.

ABSTRACT

This paper deals with the issue of the stabilization of railway track geometry. It details the published results in numerous international journals. Having analysed the cited publications the paper deals with a new research topic related to geogrid-reinforced railway ballast. A research team of the Department of Transport Infrastructure and Municipal Engineering at the Szechenyi Istvan University would like to continue working on this research topic.

Szabolcs FISCHERemail:[email protected]:toinvestigationoftheeffectofgeogridlayersunderballast.

Ferenc HORVÁTemail:[email protected]:designofrailwaysub-andsuperstruc-tures,andrailwaydiagnostics

SzéchenyiIstvánUniversity,H-9026Győr,Egyetemtér1.

Vol. XIX, 2011, No. 3, 22 – 30

1. INTRODUCTION

Geosynthethics have been used for soil reinforcement for manydecades.Theycancompensateforsomeweaksoilparameters;e.g.,theyensureadditional tensileandshearstrengthforsoilstructures.Inthiswaysoilswithunsuitablegeotechnicalparameterscanbeusedfor the constructionof steep slopes;with these types of soils veryquickconsolidationcanalsobeachievedonbadqualitysubgrades.Forexample,geosynthetic-reinforcedsoilsarealsosuitable for theconstructionofverygoodinsulationandseparationlayersatwastedumpsandhelpdecreasethewaterpermeabilityofsoils(Szepesházi,2008). Geogrids ensure additional tensile and shear strength forboth granular and cohesive soils. Using this soil reinforcement,soilscanbearnotonlypressurebutalsotensileforcetoo,duetothelongitudinal and tranversal ribs and junctions of ageogrid if thereisadequatesoil-geogridinteraction.Thisreinforcedstructureworkssimilarlytoreinforcedconcreteinwhichtheconcretebearspressingforces,andreinforcingsteelbarstakeupthetensileforces.Geogrids

canbeused for counteracting shear forces,but theiruse is limitedbecauseoftheshearstrengthofgeogrids.Laboratorytestshavebeenconducted,whicharerelatedtosuchuseofgeogridsinrailwayconstructionswherethesubgrade/subsoilwasnotreinforced,butthebed-ballastedrailwaysuperstructurewas.Inthese structures geogrids were placed under – and in some casesinto– theballastmaterial. Itwasexpected that thegeogridwouldclampparticlesofballastmaterialatthebottomoftheballast;inthiswaytherailwaytrackwouldbefloatingintheballastmaterialandexposedtodynamiceffects.Forothervibrationsthestructurewouldbemoreresistanttothedeformationofthetracks.Thisphenomenonisan interlockingeffect (Figure1),whichcanbe imaginedas theparticlesof theballastmaterial– likeeggs inaneggcarton–arewedged into the apertures of the geogrid. In thismanner aquasi-strongandrelativelyskidprooflayerwouldbeguaranteedforotherparticleslyingaboveandinterlockedintotheseparticles.The aim of this paper is to summarize the results of relevantinternational publications and, based on these, to give

2011/3 PAGES 22 — 30

23INVESTIGATIONS OF THE REINFORCEMENT AND STABILISATION EFFECT OF ...

acomprehensivereview,especiallyaboutlaboratorytestsofgeogridlayersunderballast.Theothergoalofthispaperistoformulatethetargetof theresearchon this topicat theDepartmentofTransportInfrastructure and Municipal Engineering of Szechenyi IstvanUniversity.

2. pREVIOUS INTERNATIONAL RESEARCH

International research teams have analysed geogrids for thereinforcementofrailwaysuperstructuresinthreedifferentways.Thefirstmethodwasalaboratorytestwherefull-scaleorreduced-scaleassemblieswereexaminedwithcyclicandstaticloadings.Inorder to approximately and adequately determine the prospectivebenefitsofthegeogrids,referencemeasuremenrsshouldhavebeenperformed.Thereferencemeasurementsweregenerallydonewithassembliesinwhichnogeogridreinforcementwaslaid;inthiswaytheeffectcouldbeeasilydemonstratedwith thedifferences in theresults.Section2.1dealswiththistopicindetail.Thesecondmethod involvedfield testswheregeogridswerebuiltin railway lines carrying great volumes of traffic after aballastreplacementorcleaning,andtampingwork.Thebeneficialeffectofthegeogridcanbedemonstratedwiththesettlementsinthefunctionoftheelapsedtimefromthesettinginorthenumberofaxlespassed(gross-ton). Here, reference measurements were needed, whichshowedeithertheconditionsbeforethelayingofthegeogridinthesameplaceortheconditionsofaconnectedsectionwheretherewasnogeogrid-reinforcement.The third method is very useful for decreasing the number oflaboratory testsanddecreasing theveryhighcosts related to fieldtrials;also,withthismethodtheexactnessofthemeasurementscan

bechecked.Thismethodwasacomputer-aidedsimulationwithtwopossibilities:asimulationwiththefiniteelementmethod(FEM)orasimulationwiththedistinctelementmethod(DEM).This paper mostly deals with the results published in foreignliteratureandexperiencewithlaboratorytestsconnectedtogeogridlayersbuilt underballast; it alsobriefly introduces the computer-aidedsimulationsandfieldtests.

2.1. Laboratory tests and their results

Papers dealing with geogrid-reinforced railway ballast mostlypublish the results of laboratory tests. Sections 2.1.1 and 2.1.2summarizeseveralparametersoftheselaboratorytests;Section2.2demonstratestheirresults.

2.1.1. Material, loading and data recording parameters of laboratory tests

The loadings applied during laboratory tests are an essentialparameter. In the event of modest laboratory equipment, staticloadingcanalsobeused,butstatic loadinghasmainlybeenusedduring geogrid pullout tests (with airbags built in the top sectionofabox),whichisadequateforevaluatingsoil-geogridinteraction(Nejad and Small, 2005; Perkins and Edens, 2003; Shuwang, etal., 1998).Another possible use of static loadings is to completethemeasurementswithdynamic loading (Indraratna, et al., 2006;Indraratna, et al., 2007; Raymond, 2002; Raymond and Ismail,2003). In the latter publications static loading had to be used toensure thereproducibilityof themeasurements(Raymond,2002).Inparticular,staticloadingshavetobeappliedtobeabletonoticethe failure in the caseof astatic loading situation (RaymondandIsmail,2003),andinthepapers(Indraratna,etal.,2006;Indraratna,etal.,2007),horizontalstaticloadingswereusedonmovableandloadable box walls. As arailway load is dynamic, much moreprecise resultscanbeachieved ifdynamic-cyclic loading isused.The magnitude of the cyclic loading depends on the size of theloadingplate(becauseofthestressundertheplate);thefrequency-rangeisregulatedbystandards,butanapproximatevaluefor thisrangecanbecalculatedfromthedesignvehicle’sspeedandaxle-base. In the cited publications assemblies were examined at the0-100kNload-rangeandthe0.5-15Hzfrequency-range(Bathurstand Raymond, 1987; Brown, et al., 2006; Brown, et al., 2007;Indraratna, et al., 2006; Indraratna, et al., 2007; Matharu, 1994;Raymond, 2002; Raymond and Ismail, 2003; Shin, et al., 2002;Thom, 2009) (these values are not matching parameters, but areonlyextremevalues).Oneofthemostimportantparametersfortheassembledlayersisthedepthoftheballastbed.Astheballastbed’sdepthisapproximately250-350mmintherailwaytracks(underthesleeper’slowersurface

Fig. 1 The interlocking effect (Konietzky, et al., 2004.).

24 INVESTIGATIONS OF THE REINFORCEMENT AND STABILISATION EFFECT OF ...

2011/3 PAGES 22 — 30

intheaxleoftherailatthedeterminantside),thedepthshouldbeaboutthesamevalueinanylaboratorytests.Thisdepthdependsnotonlyontheminimaltechnicaldepthofthetampingworksbutalsoon theadequateflexible load-bearing.The locationof thegeogridlayer is restricted, because thehammersof the tampingmachineswouldcatchthegeogridanddamagethereinforcedsuperstructure.In this respect the laboratory tests of papers (Raymond, 2002;RaymondandIsmail,2003)arequeriable,becausemicrogeogridswereplacedclosetothebottomplaneoftheloadingplate.Certainlythe placement ofmore than one geogrid reinforcement layer intotheballast isalsopossible,but it isaverydifficultandexpensiveundertakingwithrespecttoconstructionandmaintenance.The4%cross-slopedsubgradeplanecouldalsobemodelledinalaboratory,but it has not been mentioned in the cited publications. It isworth considering the setting in of thegeogridonto sandygravelprotection-strengthening layers,orontosub-ballast (Brown,etal.,2006;Brown,etal.,2007;Indraratna,etal.,2006;Indraratna,etal.,2007;Shin,etal.,2002;Thom,2009).Toavoidthepushingoftheballastmaterial particles into the subgrade, geotextiles should beapplied (Shin,etal.,2002) in (Brown,etal.,2006;Brown,etal.,2007;Thom,2009)combinedgeotextile-geogridlayerswereused).Themeasurementswill have amuchwider spectrum andwill bemuch more precise, if more types of parameters are applied inlaboratorytests.ThisisrelatedtoYoung’smodulusofthemodelledsubgrade or subsoil (Brown, et al., 2006; Brown, et al., 2007;Raymond,2002;RaymondandIsmail,2003;Thom,2009),andtothetypeofgeogridapplied(Brown,etal.,2006;Brown,etal.,2007;Indraratna, et al., 2007; Raymond, 2002; Shuwang, et al., 1998;Thom,2009)andtheballastbed’smaterial(Indraratna,etal.,2006;Indraratna, et al., 2007;Nejad andSmall, 2005;Raymond, 2002;Raymond and Ismail, 2003).These latter publications consideredthese three parameters with aminimum of two different values.The different Young’smoduli were considered with open-cellneoprene rubber, closed-cell pure gum amber and other rubbersheets, aswell aswith soft soils. It shouldbenoted that the casewithout ageogrid-reinforcement layer would be agood referencemeasurement.During the analysis of the ballast bedmaterial, thegrainsizeanddry-wet, fresh-recycled,clean-foulballastmaterialsshouldbeconsidered.In thepublications(Indraratna,etal.,2006;Indraratna,etal.,2007),alltheabovecaseswereconsideredexceptforthecaseofclean-foulballastmaterials.Analysingthegeogrids,themaximumtensilestrength,theyieldpointstrain,thestiffnessofthe ribsand junctions, the tensilestrengthata5%strain,aswellas the aperture size are determining parameters; when changingthese parameters, the behaviour of the geogrids can be qualifiedadequately and precisely. In the papers the required compactionmethods are completely described for all the layers (the samenumberofcompactioncycles,samecompactionwork,etc.).

Thelaboratorytestsrequireexactmeasurementdata,whichcanbereceived by precisemeasuring apparatus.These data are recordeddigitally.Settlementsduringtestswithstaticandcyclicloadingandthepulloutlengthsatpullouttestswererecordedwithlinearvariabledifferential transducers (LVDT); the forceswere recordedvia loadcells.Themeasurementsshouldberepeatedforaminimumoftwoto three times toavoidmeasuringerrors,butsuchrepetitionshavenotbeenmentioned in thecitedpublications. In thecaseofcyclicloading the number of cycles has to be registered, because thesettlementscanbeevaluatedbythefunctionofthenumberofcycles.

2.1.2. Build-up of loading structure and related parameters of the laboratory tests

Theconditionofthelaboratoryshouldbedeterminedastowhetherthetestsaredonewithfull-scale(forballastmaterialandgeogridsusableforrailwaytracks)orreducedscaleassemblies.Laboratorytestswithreducedscaleassembliesareveryquestionable(Raymond,2002;RaymondandIsmail,2003)astheycouldprovideinadequateresults.Asmicrogeogridsarenotusedintransportinfrastructures,additional laboratory tests should be done. Railway tracks havesignificant dimensions in longitudinal and transversal directionstoo. In the caseof aone-track railway line, the largest dimensionof aballast bed in across direction is approximately 4.5-5.5 m;therefore, large laboratory assemblies have to be built. Themostprecise and accurate results canbe achievedwith laboratory testsof afull-scale railway track assembly (Bathurst and Raymond,1987;Matharu,1994;Thom,2009)(Figures2-3).Onepaper(Shin,etal.,2002),where theboxwasbuiltwith1400×1000×2000mmdimensions in its length,widthandheight respectively, shouldbementioned.

Fig. 2 Modelling a full-scale railway track assembly (Matharu, 1994).

2011/3 PAGES 22 — 30


Toreducethefrictionbetweentheballastmaterialparticlesandthetestbox,careshouldbe taken that thebox issignificantlysmallerthan the width of the ballast bed in actual railway tracks. In allhonesty,itisnotasimulationofrealconditions,butitismuchbetterthan acase without any friction reduction. This solution shouldmainly be applied during geogrid pullout tests (Nejad and Small,2005;PerkinsandEdens,2003),becauseowingtotheparticlejamattheboxwalls,muchhigher(false)pulloutforcescanbemeasured.Friction reduction methods have also been mentioned duringlaboratorytestswhichmodelledareinforcedrailwaysuperstructure(Brown,etal.,2006;Brown,etal.,2007;Thom,2009).The loading plates should have such dimensions that the stressderived from the test loading applied and the area of the loadingplate will be approximately equal to the stress derived from thewheel load and the dimensions of the sleeper.The loading platesshould be made of steel or concrete for rigid behaviour, thusavoiding their influence on the results of the tests.Air bagswereapplied in (Nejad and Small, 2005; Perkins and Edens, 2003;Shuwang, et al., 1998); aluminium plate was used in (Raymond,2002;RaymondandIsmail,2003);asteelloadingplatewasappliedin(Brown,etal.,2006,Brown,etal.,2007,Thom,2009);awoodensleeperwasusedin(Indraratna,etal.,2006;Indraratna,etal.,2007),andaconcretesleeperwasappliedin(Matharu,1994;Thom,2009).Astheslumpingoftheballastbed’sshouldersduetothevibratingeffect of the railway load also generates and intensifies thedeterioration process of arailway track, this topic should behighlighted. The simulation of aballast bed’sshoulder is difficultbecause full-scaleassemblieshave tobemade,andcyclic loadingshouldbeapplied.Inthecaseofboxeswithafull-heightwall,thesimulationoftheslumpingoftheballastbed’sshoulderisimpossible

Fig. 3 Cyclic loading equipment with a full-scale railway track assembly (Thom, 2009).

Fig. 4 Triaxial test with full-height wall boxes (Indraratna, et al., 2007).

Fig. 5 Laboratory test where the ballast shoulder was built up (Brown, et al., 2007).


2011/3 PAGES 22 — 30

(Indraratna,etal.,2006; Indraratna,etal.,2007;Raymond,2002;Raymondand Ismail,2003;Shin,et al.,2002) (Figure4).This isgood as akind of approximation, but actual conditions cannot beconsideredwiththismethod.Inthepapers(BathurstandRaymond,1987; Brown, et al., 2006; Brown, et al., 2007; Matharu, 1994;Thom,2009),theballastbed’sshoulderwasbuiltup;therefore,theresults show the track stabilization effect of geogrid layers underballastmoreadequately(Figure5).

2.1.3. Results of the laboratory testsIthasbeenmentionedthatinthepapers(Raymond,2002;Raymondand Ismail, 2003), reduced scale assemblies were used for thelaboratory tests.Although the results of these measurements arequeriable, it has to be highlighted that in (Raymond, 2002),asignificant reduction of settlement (50% and 13-30%) could beachievedwithbothrounded(uncrushed)andcrushedparticles.Theaggregate crushingwas decreased due to the geogrid layers. Themost unbelievable result is that the most effective reinforcementwas measured close to the loading plate (at avery small depth)(RaymondandIsmail,2003).Thissettingisimpossiblebecauseofthetechnology;inotherrespectsitistruethattheinterlockingeffecthasthelargestvalueontheplaneofageogridlayer,butforthisyouneedenoughsoilcoveringthedepth.Intheauthors’viewtherewerenotenoughsoildepthsinthemeasurementsofthepaper(RaymondandIsmail,2003).Themeasurementsdescribedin(Shin,etal.,2002)showedthatthelargestreductionofsettlementscouldbeobtainedwiththreegeogridlayers(onewouldbebetweenthesubbaseandthesubgradesoil;onewouldbe in thesubbase,andonewouldbeunder thesubballast).Inthispaper,ahugemeasuringbox(1400×1000×2000mm),layersassembledattheirfullheight,anactualrailwaysleeperandcyclicloading were used. The most effective forming was received inthe case of combined geotextile-geogrid layers, which were setinbetween thesubbaseand thesubgradesoil (the totalsettlementreductionwas33%).Thewidest range of the research can be found in (Indraratna, etal.,2006)and(Indraratna,etal.,2007).Theresultsofthesepapersshow that the lowest plastic settlements were obtained whenusing fresh ballast materials with geogrid layers, and the largestplastic settlement were achieved with recycled ballast materialswithoutageogridreinforcement.Therewerepositiveeffectsofthegeocomposite andgeotextile reinforcements to reduce settlementsoftherecycledandfreshballastmaterials,butinthegeocompositereinforcement,onlythegeogridistheworkingpart.Aftertheinitiallarge values (until 100,000 cycles) in the settlements and in thelateralandverticalstrainsoftheballast,thesevaluesconsolidated.Thelateralstrainsoftheballastwerereducedinthecaseoffresh,dry ballast materials, but this kind of positive effect can also be

achieved in the case of recycled andwet or dry ballastmaterialsusingageotextileorageocomposite.Suchpositivereinforcementandstabilizationresultswereobtainedduring the laboratory tests using full-scale assemblies (Bathurstand Raymond, 1987; Brown, et al., 2006; Brown, et al., 2007;Matharu,1994;Thom,2009).Accordingtotheresultsof(Bathurstand Raymond, 1987) , at 39% of CBR-value 4.75 times, and at1%ofCBR-value4.9times,morecyclesgenerated25mmplasticsettlements in the case of aTensar geogrid reinforcement thanwithoutit.(Matharu,1994)showedthatgeogrid-reinforcedrailwayballast has apositive effect on the retardationof thedeteriorationprocess. The most important result in (Brown, et al., 2006;Brown,etal.,2007;Thom,2009)isthatthemosteffectiveplasticsettlement reductionwas received (considering30,000 cycles and50mm crushed particles)with geogrid apertures 1.4 times largerthantheparticlesize(Figures6-7).Forthebestreinforcement,thestiffness value of the geogridwas 1050-1150MN/m, considering30,000cycles.Theslumpoftheballastbed’sshoulderwasexactlymodelled in (Bathurst and Raymond, 1987; Brown, et al., 2006;Brown,etal.,2007;Matharu,1994;Thom,2009).

Fig. 6 Influence of the aperture size on settlements (Brown, et al., 2007).

Fig. 7 Settlements vs. number of load cycles (Brown, et al., 2006).

2011/3 PAGES 22 — 30


Inthecitedpapersrelatedtogeogridpullouttests(NejadandSmall,2005;Perkins andEdens, 2003;Shuwang, et al., 1998), there areno new scientific results. The results of the measurements andconclusions(residualshearstrength,etc.)canbefoundintechnicalbooks (Szepesházi, 2008). New research results can be found in(Perkins and Edens, 2003): in testing the pullout behaviour ofageogrid, alinear elasticmodel gives adequate results; therefore,themore difficult bounding surface plasticitymodel need not beused.

2.2. Field tests

TestsmostlyconductedinHungaryandabroadhavebeenreportedin the geogrid manufacturers’ product guides (e.g., Tensar, NaueFasertechnik, Viacon, etc.). This topic is just briefly mentionedin thispaper.The field tests are summarized in (Raymond,2002;RaymondandIsmail,2003)(Figure8),while(Brown,etal.,2006;Brown,etal.,2007;Thom,2009)givemoredetaileddescriptionsofthem.Itcanbeundeniablyagreedthatgeogridreinforcementsundertheballasthaveabeneficialeffectonthegeometricstabilizationofarailway track.Usinganadequategeogrid type, the time intervalbetween tamping works can be lengthened to 2-2.5 times thanwithout the geogrid layer; therefore, it is avery cost-effectiveand beneficial solution, because the additional construction costsofgeogrids aremuch lower than the total constructioncosts.Thesetting of geogrid reinforcement layers is suggested not only fornew railway tracks but also for old, distorted tracks.Consideringtheadditionalcostofgeogridlayersandtheincreasedmaintenancetimeinterval,therateofreturncanbecalculated;thus,remarkablenationaleconomicbenefitscanalsobementionedwhenusingthem.

2.3. A short introduction to the computer-aided simulations

Geogrid-reinforced railway superstructures can be simulatedwithfiniteelementmethodsordistinctelementmethodsintwoorthreedimensions.With the finite element method the subsoil, geogrid(s), railwayballast, loading plate and sleeper are modelled as differentelements, which are considered either as distinct “big” elementsoraredividedintoafinitenumberofpartsbygeneratinganinnernetwhichhasafinitenumberofnodes.Next,thematerialmodels,connection properties, etc., should be defined and, consideringtheseparameters,numericalmethodshavetobeusedordifferentialequationshavetobesolvedtoachieveresults.Inthecaseofdistinctelementmethods,theballastmaterialparticlesaremodelled with defined diameter balls or clumps (= complex,multiple balls). In thismethod every ball is adistinct or discreteelement(Figure9).Theassemblyoftheballsisusuallygeneratedwith the help of aparticle size distribution formodelling railwayballast,butforsomeexamples,randomgenerationisalsoacceptable.Therearedistinctelementnumericalsoftwares(PFC,OVAL,etc.),whichconsiderparticlesasrigidballs,butotherprogramscanalsobeusedwhichcalculatewithdeformableparticles(UDEC).Inthefirstgroup,thematerialpropertiesareconsideredwiththecontacts;in the second group stress-strain functions are required for thecalculations.Modelling of the geogrids can be appliedwith bothtypes(Bagi,2007).Distinct elementmethods providemore precise results than finiteelementmethods,but forDEM,prior laboratory tests areneeded.Thedistinctelementmodelhastoberefinedwiththeresultsofthelaboratory tests.The resulting computermodel shouldbehave thesame as the laboratorymodel. If it is ensured, simulating furthermeasurementsismuchsimplerthantime-consumingandexpensivelaboratoryandfieldtests.Becauseofthelimitedsizeofthispaper,(Bussert, 2009; Konietzky, et al., 2004) and (McDowell, et al.,2006)are justmentioned, the topicsofwhicharerelated toDEMsimulationsofgeogrid-reinforcedrailwayballast.

Fig. 8 Installation of a geogrid layer under a railway track ballast (Raymond, 2002).

Fig. 9 Ball and clump elements in 3D.


2011/3 PAGES 22 — 30

3. FURTHER RESEARCH IDEAS

3.1. Multi-level shear box tests

Theauthorsattempted toapproximate theeffectofgeogrid layersunder railway ballast material by measuring settlements due toverticalandhorizontalstaticanddynamicloadings,theinteractionbetween theballastmaterial andgeogrid, aswell as the reductionin aggregate crushing. In the case of ballast bedded railwaysuperstructures, the depth of the ballast material is adeterminingparameter.Thegeogridreinforcementlayerundertheballasthasanadditionalclampingeffectfortheballastmaterialonlyuptoaheight“h”. In reviewing the laboratory tests published in internationalpapers,itwasfoundthatnotestswerecompletedtodeterminethepreciseeffectofinterlockinginaverticaldirection.In(Konietzky,etal.,2004;McDowell,etal.,2006),thegeogridpullouttestsweremodelled with computer-aided simulations; the result was that inthecaseofaspecificballastmaterialandgeogrid,theinterlockingeffect of the geogrid extends approximately to a+/- 10 cm zone.Our university research team plans to do laboratory tests withamulti-level shear box to examine the change in the interlockingeffect in the function of the vertical distance upwards from thegeogrid’splane,whichhasalsobeenmissingfromtheinternationalresearchtopicstoo.Theschematicdrawingofthemulti-levelshearboxisshowninFigure10.During the laboratory tests, subgrades with three differentYoung’smoduli(e.g.,9,15and25MPa),freshandrecycledballastmaterials, vertical loadings with three different values, and threedifferenttypesofgeogridswillbeconsidered,whichmeansatleast54 cases. The vertical function of the interlocking effect of thegeogridcanbedeterminedbymeasuringtheresistanceforces(FH)duringthehorizontalshearingatfourshearplanes.Consideringthe

fourshearplanes,atleast216measurementswillhavetobedone,whichisverytime-consumingwork.Thesamecompactnessoftheballast will be guaranteed to the produce same compaction work(samelayerdepths,sametools,samecompactioncycles,etc.).TheYoung’smodulusofthematerialinthelowerframe,theparticlesizedistribution, thevaluesof theverticalandhorizontal loadings, thelateraldisplacementofthelowerframeoftheshearbox,aswellasthesettlementswillhave tobemeasuredandrecorded.Therewillbesmallwindowsontheside(paralleltotheshearingdirection)oftheupperframesandintheframeunderthegeogridlayerinordertomonitortheparticles’movements.

3.2. Field tests

Because during laboratory tests some approximations have to beapplied for modelling the actual conditions, field tests are alsoneeded.Forthesemeasurementstestsectionswillhavetobebuiltonrailwaylinescarryinghighvolumesoftraffic.Thetestsectionswillhave to contain sectionswith andwithout geogrid reinforcementsin order to have reference measurements. It is suggested thatgeogrid layers be set into existing tracks during ballast cleaningwork.Duringtheinstallationtheverticalandhorizontalpositionofthe rails shouldbemeasured at every fifth sleeperwith a0.1mmaccuracy.

Forthesemeasurements,specialsurveycolumnsorreferencescrewsin theoverhead linepoleswill have tobe applied (Figure11).At(previously) defined time intervals (1, 2, 6, 12, 18, 24 months)repeatedgeodesicmeasuringwillhavetobedone.Settlementsthatoccuringeogrid-reinforcedsectionswithdifferenttypesofgeogridscanbecomparedtoeachotherandtothesettlementsthatoccurinsectionswithout ageogrid reinforcement. In thisway the railwaytrack stabilisation effect of the geogrids and the savings due tothegeogridreinforcementcanbedetermined.Thefunctionsofthetrackdeformations(horizontalandvertical)andthetrackdistortionasafunctionof the through-rolledgross tonor thenumberof thethrough-rolledaxlescanbecalculated.Theminimalrequireddepthoftheballastcanbedeterminedfromthelaboratorytests.

Fig. 10 Schematic drawing of a multi-level shear box (shearing at shear plane No. 1).

Fig. 11 Schematic drawing of the field tests.

2011/3 PAGES 22 — 30


3.3. DEM simulation

WiththeDEMsimulationspublishedinthepapers(Bussert,2009;Konietzky, et al., 2004; McDowell, et al., 2006), the commonbehaviour of the geogrid layer and the ballast material particlescanbeeffectivelydetermined.Forthistheresultsofthelaboratorytests have to be available.Thesedatawill be used in thePFC3D(Particle Flow Code in three dimensions), which is asoftwaredeveloped by ITASCAConsultingGroup Inc. Our research teamwouldliketoverifythetime-consumingandexpensivelaboratorymeasurementswithDEMmodelling,andwewouldliketocompletethesemeasurementswithsimplecalculations.Theteamwantstouseclumpsforthecalculations.Ifagreatdealofcases(geogridtypesandnumbersoflayers,contactpropertiesoftheparticles,etc.)areconsidered,morepreciseresultscanbeobtainedfromthegeogrid-reinforcedtrackusingtheDEMsimulation.ThemovementsoftheparticlesmonitoredwithdigitalcamerasthroughthesmallwindowscanbereproducedwiththePFC3Dandcomparedtoeachother.

4. SUMMARY

The geometric stabilization effect of arailway track for geogridlayersbuiltinunderaballastcanbeverifiedwithlaboratorytests,field trials and computer-aided simulations. In the internationalpublications there are good practices and bad examples too.Allthe data dealingwith laboratory and fieldmeasurements, aswellasDEMsimulations,weretakenfromtheinternationalliterature.In (Raymond, 2002; Raymond and Ismail, 2003), reduced scaleassemblies used for laboratory tests are discussed the results ofwhich areveryqueriable; however, rounded (uncrushed)particlesaregenerallyworseforrailwayballastaggregatesthancrushedones.(RaymondandIsmail,2003)achievedamoresignificantreductionin settlement with rounded particles (50%) than with crushedones (13-30%). (Raymondand Ismail, 2003)didnot consider thetechnicalrequirements(ballastdepthovergeogrids),butitisoneofthemostimportantfactorswithrespecttoconstructionbecauseofthemaintenancework(tampingdepth).Numerouspapershavepointedoutthefactthattracksettlementsareable tobe reducedbygeogrid reinforcements.Freshand recycledballast materials and wet and dry aggregates with geogrids and/or geocomposites aswell aswithout geosynthetics (Indraratna, etal.,2006;Indraratna,etal.,2007)wereconsideredtoo.Laboratorytests made with full-scale assemblies (Bathurst and Raymond,

1987; Brown, et al., 2006; Brown, et al., 2007; Matharu, 1994;Thom, 2009) demonstrated similarly good results due to geogridreinforcements.Themostimportantresultin(Brown,etal.,2006;Brown,etal.,2007;Thom,2009)isthatthemosteffectiveplasticsettlementreductionwasachieved(considering30,000cyclesand50mmcrushedparticles)withageogridaperture1.4 times largerthan theparticle size.Theslumpof theballastbed’sshoulderhasalso beenmodelled (Bathurst andRaymond, 1987;Brown, et al.,2006;Brown,etal.,2007;Matharu,1994;Thom,2009).The results from field tests are mostly published in geogridmanufacturers’ product guides (e.g., Tensar, Naue Fasertechnik,Viacon,etc.),buttherearealsopapers(Raymond,2002;RaymondandIsmail,2003;Brown,etal.,2006;Brown,etal.,2007;Thom,2009)dealingwiththistopic.Thesecitedpublicationssummarizedthe fact that tracks constructed with geogrid-reinforced railwayballast provide better resistance against settlements. It can beundeniablyagreed thatgeogridreinforcementsunderaballastbedhaveabeneficialeffectonthegeometricstabilizationofarailwaytrack.DEM modelling is adequate for simulating geogrid-reinforcedrailway ballast. Geogrid pullout tests were modelled by PFC3Din(Konietzky,etal.,2004;McDowell,etal.,2006).Thesepapersdeterminedthatinthecaseofspecificballastmaterialandgeogrids,theinterlockingeffectofthegeogridextendsapproximatelytoa+/-10cmzone.The interlocking effect’svertical function has not been exactlyclarified and published. The research team of the Department ofTransport Infrastructure andMunicipal Engineering at SzechenyiIstvanUniversitywould like tostudy this research topic.Amulti-level shear box will be used; field tests will be performed withnumerousgeogridtypes;andDEMsimulationswillhavetobeusedforthisresearch.Cost-benefitcalculationswillalsobeapplied.

Acknowledgements

Theauthors’researchissponsoredbyHungarianRailways,TensarInternational Ltd., Gradex Ltd., Naue Fasertechnik Ltd., as wellasViaconHungary Ltd; and supported byTAMOP-4.2.1/B-09/1/KONV-2010-0003: Mobility and Environment. The Project issupported by the EU and co-financed by the European SocialFund.TheauthorswouldliketothankDénesSzekeres(HungarianRailway) for his scientific and technical help, and to KatalinKoncz(civilengineeringstudent)forherhelpintheprocessingofliterature.


2011/3 PAGES 22 — 30

REFERENCES

[1] Bagi, K. (2007): Diszkrét elemek módszere (The distinctelementmethod),universitycoursebook,BudapestUniversityof Technology and Economics, Department of StructuralMechanics,Budapest,2007,p.73(inHungarian)

[2] Bathurst,R.J.,Raymond,G.P.(1987):Geogridreinforcementofballastedtrack,TransportationResearchRecordNo.1153,1987,pp.8-14

[3] Brown, S. F., Thom,N.H., Kwan, J. (2006): Optimising thegeogrid reinforcementof rail trackballast, conference issue,RailfoundConference,Birmingham,2006,pp.346-354

[4] Brown,S.F.,Kwan,J.,Thom,N.H.(2007):Identifyingthekeyparameters that influence geogrid reinforcement of railwayballast, Geotextiles and Geomembranes, Vol. 25, 2007, pp.326-335

[5] Bussert,F.(2009):Recentresearchintotheactualbehaviourofgeogridsinreinforcedsoil,conferencerepresentation,JubileeSymposium on PolymerGeogrid Reinforcement, September8,2009,London

[6] Indraratna, B., Shahin, M., Rujikiatkamjorn, C., Christie, D.(2006): Stabilisation of ballasted rail tracks and underlyingsoftformationsoilswithgeosyntheticgridsanddrains,ASCESpecialGeotechnicalPublicationNo.152,ProceedingsofGeo-Shanghai2006,2-4June,2006,Shanghai,China,pp.143-152

[7] Indraratna,B.Shahin,M.A.,Salim,W.(2007):Stabilisationofgranular media and formation soil using geosynthetics withspecial reference to railway engineering, Journal of GroundImprovement,Vol.11,No.1,2007,pp.27-44

[8] Konietzky, H., Te Kamp, L., Groeger, T., Jenner, C. (2004):Use of DEM to model the interlocking effect of geogridsunder static and cyclic loading, Numerical Modeling inMicromechanicsViaParticleMethods—2004(Proceedingsofthe2ndInternationalPFCSymposium,October2004,Kyoto,Japan),Y.Shimizu,etal.,eds.Leiden:Balkema,2004,pp.3-11

[9]McDowell, G. R., Konietzky, H., Jenner, C., Harireche, O.,Brown,S.F.,Thom,N.H.(2006):Discreteelementmodellingof geogrid-reinforced aggregates, Geotechnical engineering,Vol.159,No.1,2006,pp.35-48

[10] Matharu,M. (1994):Geogrids cut ballast settlement rate onsoft substructures, Railway Gazette International, Vol. 3,March1994,pp.165-166

[11] Nejad,F.M.,Small,J.C.(2005):Pulloutbehaviourofgeogrids,Iranian Journal of Science & Technology, Transaction B,Engineering,Vol.29,No.B3,2005,pp.301-310

[12] PerkinsS.W.,EdensM.Q.(2003):Finiteelementmodelingof ageosynthetic pullout test, Geotechnical and GeologicalEngineering,Vol.21,2003,pp.357-375

[13] Raymond,G.P.(2002):Reinforcedballastbahavioursubjectedto repeated load, Geotextiles and Geomembranes, Vol. 20,2002,pp.39-61

[14] Raymond G., Ismail, I. (2003): The effect of geogridreinforcement on unbound aggregates, Geotextiles andGeomembranes,Vol.21,2003,pp.355-380

[15] Shin,E.C.,Kim,D.H.,Das,B.M.(2002):Geogrid-reinforcedrailroadbed settlementdue to cyclic load,Geotechnical andGeologicalEngineering,Vol.20,2002,pp.261-271

[16] Shuwang,Y.,Shouzhong,F.,Barr,B. (1998):Finite-elementmodellingofsoil-geogridinteractiondealingwiththepulloutbehaviour of geogrids, Acta Mechanica Sinica (EnglishSeries),Vol.14,No.4,Nov.1998,pp.371-382

[17] Szepesházi,R.(2008):Geotechnika(Geotechnics),universitycoursebook,SzechenyiIstvanUniversity,Gyor,2008,p.187(inHungarian)

[18] Thom, N. H. (2009): Rail trafficing testing, conferencerepresentation, Jubilee Symposium on Polymer GeogridReinforcement,September8,2009,London

SZ. FISCHER, F. HORVÁT Szabolcs FISCHER Ferenc HORVÁT … · 22 2011 slovak university of technology sz. fischer, f. horvÁt investigations of the reinforcement and stabilisation

Documents