Experimental study of sandwich structures as armour against medium-velocity impacts Amélie Kolopp, Samuel Rivallant, Christophe Bouvet To cite this version: Amélie Kolopp, Samuel Rivallant, Christophe Bouvet. Experimental study of sandwich structures as armour against medium-velocity impacts. International Journal of Impact Engineering, 2013, vol. 61, pp. 24-35. 10.1016/j.ijimpeng.2013.05.007. hal-00906305 HAL Id: hal-00906305 https://hal.archives-ouvertes.fr/hal-00906305 Submitted on 19 Nov 2013 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Open Archive TOULOUSE Archive Ouverte (OATAO) OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible. This is an author-deposited version published in : http://oatao.univ-toulouse.fr/ Eprints ID : 9325 To link to this article : DOI : 10.1016/j.ijimpeng.2013.05.007 URL : http://dx.doi.org/10.1016/j.ijimpeng.2013.05.007 To cite this version : Kolopp, Amélie and Rivallant, Samuel and Bouvet, Christophe Experimental study of sandwich structures as armour against medium-velocity impacts. (2013) International Journal of Impact Engineering, vol. 61 . pp. 24-35. ISSN 0734-743X Any correspondance concerning this service should be sent to the repository administrator: [email protected] Experimental study of sandwich structures as armour against medium-velocity impacts Amélie Kolopp, Samuel Rivallant*, Christophe Bouvet UniversitédeToulouse,INSA,UPS,MinesAlbi,ISAE,ICA(InstitutClémentAder),ISAE(InstitutSupérieurdel’Aéronautiqueetdel’Espace), 10AvenueEdouardBelin,BP54032,31055ToulouseCedex4,France a b s t r a c t Anexperimentalimpactstudyhasbeenconductedonsandwichstructurestoidentifyandimprovear- moursolutionsforaeronauticalapplications.Theobjectivesaretofindthebestconfigurations,i.e.the non-perforated targets with the minimal weight and back deformations. Medium-velocity impacts (120m/s)havebeenconductedusinga127gsphericalprojectile.Thetargetsaresimplysupportedatthe rearofthestructure.Twopotentialchoicesoffrontskinhavebeenidentifiedforthesandwichstructure: 3mmthickAA5086-H111aluminiumplatesanddryaramidstitchedfabrics(between8and18plies). Thedrystitchedfabricsappeartobeanoriginalsolution,whichassociatesalightweightstructureanda Keywords: goodperforationresistance.Moreover,astrongcouplinghasbeenfoundbetweenthefrontskinandthe Mediumvelocityimpact Sandwichstructure core.Theimpacttestsindicatethataluminiumhoneycombcoreassociatedwithaluminiumskinsshow Armour mitigatedresults.However,thecombinationofdryfabricfrontskinandaluminiumhoneycombshow Aluminium betterperformancesthanaluminiumsandwiches,withaglobalweightdecrease. Fabrics 1. Introduction against shocks in naval structures, or against impacts [1e3]. Sandwich structures using composite materials are foreseen as Aeronautical structures and especially zones of the aircraft armoursolutions,duetotheirstiffnessandlightweightproperties. fuselagecanbepotentiallysubjectedtoforeignobjectimpacts,like Either associations of dry fabrics or interlock structures and ice,enginedebrisorbirdstriking.Thisstudyaimstoidentifyandto ceramiclayers,eithermetallicstructures(mostlysteelthickplates) optimize armoured structures against hard projectiles through areoftenusedinballisticstudies.Amongthesematerials,ceramics experimental impact tests. These impacts are characterized by andsteelarenotchosenbecausetheirhighdensitiesdonotfitto medium-velocities and high energies: 120 m/s and 1 kJ. In this theaeronauticalconstraints.Thus,aluminium,compositesanddry paper,severalcriteriaarechosentoselectandcomparethearmour fabricsareidentifiedaspotentialmaterialsandstructuresconsid- solutions: minimal weight and residual deformation without eringmedium-velocityimpacts.Therefore,theliteraturerevuefo- perforating. Sandwich structures appear to be potential armour cusesonsandwichorlayeredstructuresusingthesematerials. architectures,providinganincreaseinbendingrigiditywithouta Considering aluminium plates, the literature relates that significantincreaseinstructuralweight.Theyarecurrentlyusedin aluminiumductilitycanplayamajorroleinimpactresistance.For numerous applications like for instance helicopter blades, ship instance,Børviketal.[4]showedthattheballisticlimitofAA5083- hullsoropticalbenchesforspaceapplications.Acompletelitera- H116 is 20% higher that the AA7075-T651 (244 m/s impact on ture review on sandwich structures subjected to impact is very 20mmthicktargets),whichisalthoughmoreresistantwithayield difficulttoestablishduetothewidevarietyoftargetmaterialsand stresstwiceashigh.Severalmaterialandthicknessescombinations impactconditions. havebeentestedintheliterature.Inballisticimpacts,theassoci- Aluminiumsandwichescomposedbyaluminiumplatesasskins ationofaductilematerialasthefirstlayerassociatedwithahigh and aluminium cores (honeycomb or foam) are currently used strengthmaterialgivesthebestimpactperformances[5](flatand conical projectiles of 200 g launched at 400 m/s on steel layers targets). The layering of aluminium or steel plates to improve impactperformanceshasbeenwidelystudiedintheliterature[6e * Correspondingauthor.Tel.:þ33561338158;fax:þ33561338352. 9].However,itisdifficulttoestablishacleartendency.Forinstance, E-mail addresses: [email protected] (A. Kolopp), [email protected] (S.Rivallant),[email protected](C.Bouvet). Guptaetal.[6]showedthat1100-H12aluminiumplatesperforma http://dx.doi.org/10.1016/j.ijimpeng.2013.05.007 betterimpactresistance thanlayeredtargetswith thesame total aluminiumhoneycomb20mmthick(velocitiesintherangeof92e thickness (1e3 mm). Nevertheless, stratification seems to be ad- 548m/s).Theskinsareidentifiedasthemainfactorresponsiblefor vantageouswiththicktargets,typicallyaround6mm[7].Marrom theenergyabsorption(respectively46%,13%and41%forthefront etal.[10]aswellasRadinetal.[11]comparedtheperformancesof skin,thecoreandthebackskin). monolithicstructuresandoflayersincontactorseparatedbyanair The previous synthesis showed that many studies were con- gap. They both showed that layered structures in contact give ducted on specific materials or structures like aluminium plates, betterresultsthanmonolithicstructuresorlayeredstructureswith dry fabric assemblies, etc. However, the interaction of different agapbetweeneachlayer,forasametotalweight. materials assembled in a structure is rarelyaddressed due tothe Concerning the dry fabrics behaviour, Cheeseman et al. [12] largenumberofexperiments necessarytoidentifycouplings and made a synthetic review on the different parameters influencing structural effects. Moreover, contrary to ballistic or low-velocity the ballistic impact performances of these structures. Theimpact impacts, medium-velocity impacts with high energies are few conditionsastheprojectilevelocity,thetargetdimension,andthe studiedintheliterature.Thisstudyaimstoidentifyandcompare boundaryconditions(aperturesize[13],targetdimension)arethe sandwich structures subjected to medium-velocity impacts. mostsignificantparameters.Forinstance,Zengetal.[14]showed Severalassembliesofskinsandcorearestudiedtodeterminethe thattheenergyabsorbedbytwoedgesclampedTwarontargetsis respectiveroleofeachpartofthesandwichandpossiblecouplings 4.5timestheabsorbedenergyinthecaseoftotalclamping(under as well as to propose ways of material and geometrical 350 m/s). The fabric orientation at 45" shows good results by optimizations. increasingtheloadingareaandthelengthof theprincipalyarns. Theexperimentalimpactset-up,targetdescriptionandimpact The fundamental parameters related to the fabric are the fibre results are given in Section 2. Then, the behaviour of aluminium material,theplynumberandthefrictionproperties.Fabricsmade sandwiches is described in Section 3, followed by the study of of aramid (Kevlar 49, Twaron), polyethylene (Spectra, Dyneema), sandwich structures with dry fabric front skin in Section 4. The PBOfibres(Zylon)aremostlyused.Roylanceetal.[15]showedthat non-perforated structures are compared in Section 5. Finally, thecoverfactorisalsoanimportantparameter.Itconsistsinthe concludingremarksaregiveninSection6. percentageofareacoveredbythefabric(relatedtothewidthand pitch of warp and weft yarns) and an optimal range is defined 2. Experimentalset-up between 60 and 95% to avoid yarn sliding under the projectile. Moreover,theyarnsurfacepropertiesarealsotobeconsideredas 2.1. Impacttestconditions frictionleadstolargeramountsofabsorbedenergy.Astudycon- ducted by Briscoe and Motamedi [16] noted that a lubricant Normalimpacttestsareconductedusingagasgun.Aspherical addiction lead to a decrease in the fabric absorbed energy. Ac- projectileof127gand30mmdiameterislaunchedatanaverage cording to the authors, the role of yarn friction in the resistance velocityViniof120m/satthecentreofthetargets.Theprojectileis againstperforationandenergyabsorptionisnottotallyunderstood composed by a hardened steel spherical nose and shank and is yetandneedsfurtherinvestigations.Karahanetal.[17]andAhmad supposedperfectlyrigid.Thecylindricalshankwith8mmdiameter etal.[18]studiedtheeffectonplynumberandstitchingonaramid and 50 mm length is screwed to the rear of the nose. The same fabrics.Karahanetal.[17]showedthatnosignificantdifferencein projectileisreusedforallimpacttests.Ahighspeedcameraisused energy absorption is observed varying the stitching pattern tomeasuretheprojectiledisplacementandvelocityduringthetest (perimeter, 50 mm edgegrids, etc).Ahmad etal. [18] showedan by following to the painted shank. Square targets of 200, 300 or increaseinballisticlimitwithstitchedtargets(þ8%with2inedges 400-mmsidearesimplysupportedattherearbyasquareframe stitching compared to targets without stitching). Note that with an aperture of 170-mm side (see Fig. 1). These boundary numerous experimental and numerical studies are conducted on conditions are more representative of impact on real structures dryfabricstructures.However,multi-layeredstructures,orfabrics such as the aircraft fuselage compared to clamping along four associated with other materials are rarely investigated in the edges. The experimental set-up and several camera pictures are literature.Moreover,theuseofdrystitchedfabricswithinasand- giveninFig.1. wichstructureisnotyetstudiedtoourknowledge. Intheparticularcaseofdryfabrics,theboundaryconditionsare Many studies have been conducted on sandwich structures definedinordertoberepresentativeofarealstructureofabout600- subjectedtoimpact[19e22].Attheonsetofimpact,compressive mmside.Duringtheimpact,theprimaryyarns(i.e.theyarnssitu- waves propagate under the projectile and provoke local out-of- ated underthe projectile) areloadedin tension, whichinduces a plane shear damage of the skins and core crushing. In a second yarn elongation. The additional distance is called de-crimping step, a global shear/bending deformation of the structure is length(seeFig.2aandb).Thisdistancecanbecalculatedknowing observed.Notethatiftheinitialprojectilevelocityissuperiortothe thestructuresizeandtheweaveproperties.Whenconsideringthe ballisticlimit,thebendingeffectdoesnotappear.Abrate[23]noted fabricsusedinthisstudy(cf.Table1)theyarncrimpreaches0.4%for that few studies worked on the sandwich configurations while thetwillfabricandrespectively0.51%and3.5%fortheplainweavein manyofthemconsideredtheprojectileparameterseffects(shape, theweftandwarpdirection.Therefore,thede-crimpinglengthcan mass and velocity). This is why results are sometimes in conflict reach2.4mminthetwillfabricandrespectively3.1and21mmin fromastudytoanother.However,itisnecessarytostudyawide theweftandwarpdirectionoftheplainfabricfora600-mmside number of sandwich configurations in order to understand the panel.Inordertorepresentthismechanismwithsmallersamples complexcouplingbetweenfacingandcorewhichlargelydictates comprisedbetween200and300-mmside,afree-edgesboundary the impact damage for a given loading. Abrate noted that the conditionhasbeenchosen.Thus,theyarnscanpotentiallyslidefrom penetrationresistanceismostlygovernedbytheoverallrigidityof the target extremity, providing an additional length under the thetargetsandthefacingpenetrationresistance.Incaseofcom- impactpoint(Fig.2candd). positefacingsandwichesfailthroughmatrixcracks,fibrefracture and delamination. The structures also exhibit core crushing and 2.2. Materialdescriptionandtargetsidentification facesheet debonding. Buitrago et al. [24] conducted impact tests and numerical simulations on sandwiches using composite face- Thetestedsandwichstructuresareassembledusingaluminium sheets (carbon fibre with epoxy resin, 2 mm thick) and an ordryfabricfrontskins.Honeycombcoreofdifferentthicknesses Fig.1. Experimentalset-upandhighspeedcamerapictures. (from 10to30mm)and materials (aluminiumandNomex) have SkinsandcoreareassembledusingaRedux!adhesivefilm.In been selected. Impact results on aluminium plates are used as the case of dry fabric front skin, the plies are stitched and then reference cases.Among thewidechoice ofaluminium alloys,the bondedtothecore.Therefore,thelastply(plyincontactwiththe ! 5XXX family alloys is chosen due to their ductile behaviour and core)istotallyimpregnatedwiththeRedux bond. their positive strain rate sensitivity which increases the skin resistance against the perforation [25]. More particularly, the 2.3. Impactresults AA5086-H111aluminiumalloyhasbeenusedasfrontand/orrear skin.Then,sandwichstructuresareassembledandtestedusingthe Theinitialvelocitypresentsdispersionsduetotheexperimental sameskinconfigurations,withtheadditionofaluminiumhoney- set-up, so the value can be substantially different between the combcoresofseveralthicknesses.Afterthat,dryfabricsaretested tested samples. However, the initial projectile velocity Vini is asanewfrontskininsandwichstructures. measuredfromthecamerapicturesforeachconfiguration,sothe Thematerialsandstructuresusedforthesandwichesassembly variability is known and taken into account in the analysis. The aredescribedinTable1. indexesRfandRrarerespectivelyassociatedwiththestatusofthe Thesampleidentificationcontainsdataaboutthefrontskin,the frontandtherearskinafterimpactandaresetto“N”,“C”or“Y”in coreandtherearskinrespectively.Eachmaterialisassociatedwith case of non perforation, limit case and perforation. The residual aletter(seeTable1)andanindexindicatesthelayerthicknessor projectilevelocityVresismeasuredaftertheendofcontactbetween the number of plies (for fabric skins). A global index reports the the projectile and the target. Positive values of residual velocity targetdimension(200,300or400-mmside).Forinstance,a200- indicateperforationwhilenegativevaluesaremeasuredduringthe mmsidesandwichcomposedbya12pliesoftwillfabricasfront projectilerebound.ThemeasureofViniandVresgivesrespectively skin, a 10-mm thick aluminium honeycomb and a 1-mm thick theinitialkineticenergyoftheprojectileEiniandresidualenergy aluminiumrearskinisnoted[T12AH10A1]200.Ifatestedstructureis Eres, which indicates the energy absorbed by the target Eabs. The assembled without core, a synthetic notation is used: a 300-mm maximumprojectiledisplacementdmaxandtheresidualindenta- sidealuminiumtargetwitha2-mmthickfrontskinanda1-mm tionIresaremeasuredinnon-perforatedcasesonly.Themaximum thickrearskinisrepresentedby[A2þ1]300. projectiledisplacementisobtainedfromthecamerapicturesdur- Additional data are given concerning the sandwich assembly. ingtheimpact.Theresidualindentationismeasuredafterimpact Fabrics areassembledaccording toaquasi-isotropicstratification by3Dcorrelationontherearfaceoftherearskin.Theweighper by combining plies oriented in 0/90" and #45" directions in the unitsurfacersisgivenforeachconfigurationinTable2. followingorder:[0"/45"]n,symwithn¼2,3,4.Thesestructuresare stitched in order to maintain the dry plies together. Therefore, a Table1 para-aramidfilamentisused(K-Tech75,0.23mmdiameter).The Skinsandcoresrepresentationandmaterialdata. stitchingpatternisasquaregridof10or20-mmside. Figure Name Description At AA5086aluminiumplates(2700kg/m3); t¼1or2mmthick. Ti Aramidtwillfabric:2/2twill,220g/m2, 6.5ends/cm.Twaron2200HMfibre (1620dtex);numberofpliesi¼8,12or16. Pi Aramidplainweavefabric:170g/m2, 6.7ends/cm.Twaron2200fibre(1210dtex); numberofpliesi¼12or18. AHt Aluminiumhoneycomb(Hexcel):30-mm thick(3/8-5052-.004,86.5kg/m3),20mm thick(ACG-3/8,53kg/m3),10mmthick (ACG-1/4,72kg/m3) Fig.2. a)Realstructureeinitialstate;b)duringimpact;c)testedtargets(freeedges NHt Nthoicmke(xHhRoHn-e7y8c-o1m/8b-3(.H0,e4x8cekl)g:/m203)mm boundarycondition)einitialstate;d)duringimpact. 3. Impactonaluminiumsandwichesandcorethickness deformations(whichimpliesagreateramountofabsorbedenergy). influence The increase of target maximum indentation is also noticed: for instance, the [A2þ2]200 target is more indented than [A2þ1]400 The study firstly focuses on “aluminium sandwiches”, i.e. despiteahighertotalthickness.Itappearsfromtheseobservations aluminiumskinswith aluminiumhoneycombcore.Severalsand- that 200-mm side targets behaviour is not representative of real wichstructureshavebeenimpactedvaryingtherearskinandthe structures.Onthecontrary,300and400-mmsidesandwichesare corethicknesses.TheimpactresultsaresynthesizedinTable2for not deformed until the target size due to inertial effects which sandwich structures (group 2) and compared with aluminium prevent the momentum of the structure near the boundarycon- plates(group1). dition. Thus, 300 and 400-mm side targets results can be easily compared, but the comparison between 200 and 300-mm side 3.1. Targetsizeeffect samplesismoredelicate. The target dimensions are suspected to influence the impact 3.2. Damagemechanismsobservations results of both aluminium plates and sandwiches. Particularly, a significantdifferenceisnoticedbetweenthe200and300-mmside Theimpactresultsonaluminiumplatesunderlinealimitrupture samples.DeformedtargetsafterimpactarepresentedinFig.3. caseforthe[A2þ1]400targetwithacircularruptureattherearofthe ItcanbeshownfromFig.3cthatresidual deformationsreach target,extendedonlytothe1-mmthickplate(seeFig.4aandb).The the plate boundary in the case of 200-mm side sandwiches. A [A2þ2]200targetisnotperforatedbutaneckingzonecanbeclearly globalbendingeffectofthestructureandthetargetrotationatthe seenneartheimpactpointatthebackofnon-perforatedsamples vicinity of the support induce significant plate and core (Fig. 4c). Aluminium plates in sandwich structures generally fail Table2 Structureconfigurationsandimpactresultssynthesis. ID Figure rs[kg/m2] Vini[m/s] Rf/Rr Vres[m/s] Eabs[J]/%Eini dmax[mm] Imax[mm] Group1:aluminiumplates [A2þ1]400 8.19 124.4 N/C %22.4 951/96.8% 41.6 30.3 [A2þ2]200 10.9 126.6 N/N %9.3 1011/99% 31.9 31.7 Group2:aluminiumsandwiches [A2AH10A2]200 11.8 120.0 N/N ea ea ea 24.7 [A2AH20A2(1)]200 12.0 122.5 C/N %6.3 949/99.5% 48.5 23.8 [A2AH20A2(2)]200 12.0 125.3 Y/Y 25.4 955/95.9% / / [A2AH10A1]300 9.0 114.9 Y/Y 62.6 590/70.3% / / [A2AH20A1]300 9.4 117.9 Y/Y 62.2 638/72.2% / / Group3a:dryfabricfrontskinsandwiches(200-mmsidetargets) [T16AH10A2]200 10.2 123.2 N/N %12.0 955/99.1% 54.4 39.3 [T8AH10A2]200 8.3 126.0 C/Y 0.0 1199b/99.9% / / Group3b:dryfabricfrontskinsandwiches(300-mmsidetargets) [T12AH30A1]300 8.4 115.2 N/N %3.0 843/99.9% 56.7 25.3 [P12AH30A1]300 8.0 123.8 N/N %5.8 971/99.8% 54.3 29.4 [P18NH20A1]300 7.5 114.4 N/N %12.9 821/98.7% 56.2 28.5 a Measureproblemoccurringduringthetest. b Particularcasewithaprojectileweightof151ginsteadof127g. onsetofimpactandtheprojectilebrakingisdecreasingafterthis limit.Apreviousstudyonaluminiumplatesshowedthatthefirst stage is attributed to the local indentation of the plates and the second is due to a structural effect (see Fig. 7). This behaviour consistsinaglobalbendingofthetargetinthesamedirectionthan the projectile (which explains the decrease of projectile braking) andoccurswhenthedeformationsintheplatesreachtheboundary condition (see Figs. 7 and 8a). This mechanism leads to a large amountofabsorbedenergy.Itisnotorpartiallyseeninperforated cases,dependingonthefrontskinruptureapparitionintheimpact sequence. The sandwich structures appear to brake the projectile less efficientlythantheskinsconfiguration,asshowninFig.6a(gapof about15m/sfrom0.16msimpactdurationbetween[A2þ2]200and [A2AH20A2(1)]200). However, the skins configuration [A2þ2]200 is more indented than the sandwich structures as shown in Fig. 6b (þ27% and þ41% indentation compared to respectively [A2AH20A2(1)]200 and [A2AH10A2]200). The [A2AH20A2(1)]200 and [A2AH10A2]200 residual profiles are veryclose and showa double curvature,globalandlocalasdescribedinRefs.[19e22].However, thefrontskinlocalcurvatureappearstobelargerinthesandwich configurationthanintheskinsalone,whichisprobablyduetothe radialslidingofthehoneycombprovidingagapbetweenthetwo skins (see Figs. 5a and b, and 6b as well as the corresponding schemeinFig.8bandc). Inthesandwichstructure,apartialde-couplingoccursduetothe separationofthetwoskins.Itconsistsinadeformationofthefront skin alone, followed by the reaction of both skins together (see Fig. 8b). This case corresponds typically to the [A2AH10A2]200 Fig. 3. [A2AH20A1]300 target: a) front view; b) rear view; c) front view of configuration where no skins rupture occurs and the plastic de- [A2AH20A2(1)]200. formationsreachtheboundaryconditionforbothskins(seeFig.5c). In a virtual sandwich configuration with a very thick core, it throughpluginitiation(Fig.4b)andtheformationofpetals(Fig.3b), couldbeenvisagedthatthefrontskinfailsbeforethecontactbe- whichisatypicalfailuremodeofductiletargets[26]. tweenbothskinsand therearskindeformation(as illustrated in Two similar sandwiches namely [A2AH20A2(1)]200 and Fig. 7c). In this case, the two skins would react simultaneously, [A2AH20A2(2)]200havebeenimpactedattwodifferentinitialveloc- resultinginadecreaseoftheenergyabsorptioninthetarget,(no ities, respectively 122.5 and 125.3 m/s. The first case structuraleffectoccurringandearlierruptureinitiation)andthus [A2AH20A2(1)]200isidentifiedasalimitcaseasonlythefrontskinis higherresidualvelocitiesoftheprojectile.Both[A2AH20A2(1)]200and perforated (Fig. 5a). The second sample is totally perforated [A2AH20A2(2)]200configurationsareintermediatecasesbetweenthe (Fig.5b).Differentdamagemechanismscanbeobservedfromre- partialandtotalde-couplingshowninFig.8bandc.Theprojectile sidualsandwichprofilesshowninFig.5.The[A2AH20A2(1)]200target velocity evolution is similar for these two configurations until showsa16-mmdiametercircularruptureoffrontskinunderthe 0.34 ms. After this moment, the two curves gradually differ, sphericalnoseandsecondarycrackspropagationresultinginpetals resultinginanegativeresidualvelocityinthefirstcase(projectile formation. In the second case [A2AH20A2(2)]200, a circular cap of rebound) and a positive value (target perforation) in the second 18 mm diameter is formed and remains partiallyattached tothe case.Consideringtheexperimentaldataavailable,noexplanation frontskin.Honeycombbucklingisobservedinbothcasesunderthe canbegiventojustifythedifferenceinthevelocityevolutionbe- front skin and the honeycomb is totally crushed under the pro- tweenthe[A2AH20A2(1)]200andthe[A2AH20A2(2)]200casesontheone jectile.Inaddition,out-of-planesheardeformationsareobservedin hand,andbetweenthesesandwichandthe[A2þ2]200skinsonthe thecorefromtheimpactpointtothetargetextremity.Highout-of- other hand. However, a potential scenario can be proposed. planeskindeformations neartheimpactleadtoaruptureof the Considering that the front skin local curvature and deformation bondbetweenthecoreandtherearskininazoneof60and85mm (dependingonthecoreheightasshowninFig.6b)ishigherinthe radius for [A2AH20A2(1)]200 and [A2AH20A2(2)]200 respectively (note sandwichstructure,theonsetofrupturemayappearearlierinthe thatinthesecondcasetherearplateisalmostseparatedfromthe sandwichthanintheplates(andprobablybeforethecontactofthe restofthetarget). twoskinsasshowninFig.8c).Therefore,theruptureinitiationand The[A2AH10A1]300and[A2AH20A1]300aretotallyperforateddue propagation in the front skin sandwich structure could be the tothetargetsizeincrease(from200to300-mmside)andtherear explanation for the first divergence between the skins and the plate thickness decrease (from 2 to 1-mm thick). However the sandwichcases,occurringafter0.2ms. correspondingreferenceofaluminiumplates[A2þ1]400isnottotally perforated(limitcaseofaluminiumplatesalone). 4. Impactofsandwicheswithdryfabricskins 3.3. Influenceofcoreinaluminiumsandwichstructuressubjected A preliminary study of the front skin choice in the sandwich toimpact impact resistance has been conducted. Dry stitched fabrics, com- positeskinsandaluminiumskinshavebeentestedasfrontskinin Asforastheprojectileevolutionofthereferencecase[A2þ2]200, severalsandwichstructures.Thecompositefrontskinsandwiches achangecanbeobservedintheslopefromabout0.4msafterthe were systematically perforated, for a similar or superior weight Fig.4. [A2þ1]400target:a)frontskin;b)rearskin;c)rearskinofthe[A2þ2]200sample. thanaluminiumskins.Thisresultisattributedtothefibre-matrix adhesion which stops the relative move of the yarns and results in a stiff but fragile behaviour. Moreover, the front skin damage remains local which implies a decrease of energyabsorption. On thecontrary,thedryfrontskinsampleswerenotperforateddespite a lower area density than the impregnated cases and no fibre rupturehas beenobserved.This result shows that the use of dry fabric skins is more advantageous than composites in medium- velocityimpacts. Thus,severalsandwichstructureswithdryfabricsfrontskinand aluminiumhoneycombandrearplatesaretestedandcomparedin thissection.Severalparametersarestudied:targetsize,numberof plies, weave pattern (twill and plain) and core characteristics (thickness and properties). The impact results of dry fabric front skinstructuresaresynthesizedinTable2(group3). 4.1. Damagemechanismsobservationofgroup3atargets The[T8AH10A2]200targetcomposedby8dryfabricpliesasfront skinistheonlyperforatedcaseofthegroup3.Notethatthesame structure with 16 dry fabric plies is not perforated. Yarn sliding appearstobeoneofthemostsignificantdamagemechanismsfor thefirstplies.Acrossshapeslidingisnoticedfortheprimaryyarns, i.e. yarns situated directly under the projectile. The observed damagearerepresentedbyaschemeinFig.9for0/90" and#45" orientedpliesandtheexampleof[T8AH10A2]200isgiveninFig.10a. Inthiscase,yarnslidingisvisiblefromthetargetboundarytothe Fig.5. Residualprofilesofdifferenttargets:a)[A2AH20A2(1)]200;b)[A2AH20A2(2)]200;c) impactlocation.Notethattheslidingdistance isverysimilarbe- [A2AH10A2]200. tweentheweftandthewarpdirectionfortwillandplainfabrics. Fig.6. Aluminiumsandwichestargets:a)projectilevelocity;b)residualprofiles. Fig.8. Propositionsforimpactscenariionskinsandsandwichstructures. In order to obtain a better understanding of the energy ab- sorption mechanisms in dry fabric skins, damage mechanisms of eachplyhavebeenobservedafterimpact. The average sliding distance of the primary yarns reach respectively 140, 70and 20 mm for the first 0/90" oriented plies (plies n"1, 3 and 5 respectively) and 45 and 0 mm for the #45" orientedlayers(pliesn"2,4).TheseyarnsarevisibleinFig.10bat the rear of the target and form a pocket which stopped the pro- jectile. The difference of sliding distance between 0/90" oriented layers and #45" layers can be associated with the superior yarn lengthandtheyarncrossingincrease. From the 6th layer, no yarn sliding is observed. The stitched points vicinities are partially impregnated with resin due to the adhesivefilmusedtobondthecoreandtherearofthefabricskin (seeFig.10c).Moreover,yarnandmatrixrupturescanbenoticed near the impact location. As a result, considering that the resin impregnation highly influences the fabric properties and impact performances(asshownintheintroductionofSection4),onlythe 5firstlayershavetobeconsideredasdryfabrics. Inadditiontoprincipalyarnsliding,otherdamagemechanisms can be observed in the front skin. Firstly, no yarn rupture is observedindryplies.Itisduetothefreeedgesboundarycondi- Fig.7. Localandglobaldeformationinasandwichstructureduringtheimpact[27]. tions and the possibility for yarns to slide instead of break. Fig.9. Damageoffabrics:a)in0/90"plies;b)in#45"plies. However,stitchingpointsarebrokenaroundtheimpactpointina zoneofapproximately60mmradius.Thisrupturemaybeassoci- ated with high out-of-plane deformations of the initial woven Fig.10. [T8AH10A2]200sandwichafterimpact:a)plyn"1ofthefrontskin;b)sideview; pattern near the impact (see Fig. 9a). However, the partial c)6thplyofthefrontskin. impregnationofthestitchpointsnearthe6thplywhichisalsoa possiblecauseofstitchingrupture. ThesamedamagemechanismsthanthosedescribedinSection 4.2. Damagemechanismsobservationofgroup3btargets 3areobservedinthecoreandtherearskinofthesandwich:partial cap formation of the aluminium rear skin (Fig. 10b), secondary Onthecontrarytoaluminiumsandwiches,noneofthe300-mm cracks,honeycomb crushingand adhesivebond rupturebetween sidesandwicheswithdryfabricskinsareperforated(Table2).The the core and the rear skin at the impact vicinity. Several energy primaryyarnslidingdistanceismeasuredfromthesamplesafter absorption mechanisms can be identified from the [T8AH10A2]200 impact. The average values reach respectively 10 and 13.8 mm target. Firstly, the wide plastic deformations of the rear skin, (Fig.11d)forplainandtwillwovenpattern,whichisintheallow- extended from the impact location to the boundary condition as able range considering the boundary conditions hypothesis (see well as the plate rupture are associated with significant energy Section2.1). absorption.Thepartofenergyabsorbedinthecoreissupposedto ThetargetobservationofFig.11aandbshowshighlydeformed be low because of the limited zone concerned. Several potential folds from the impact point to the target extremity. They are energyabsorptionmechanismsindryfabricscanbeidentified.The initiated from the last impregnated ply (the 8th ply which is in most significant one may be associated with frictional contact contactwiththecore)whichishighlydeformed.Inthesezones,a propertiesbetweentheyarnsthroughseveralcontacts:weft/warp localdebondingoftheadhesivefilmisobservedbetweenthefabric yarn crossings, contacts between the plies and between the pro- skinandthecore.Thefoldsaremorepronouncedforplainthanfor jectileandthefirstply[22]. twill weave because of the higher stiffness of the plain woven Notethattheprimaryyarnsslidingof[T8AH10A2]200issuperior pattern.Notethattheyarenotvisibleinthe200-mmsidetargets tothemaximumallowableslidingdisplacementdefinedinSection due to a global bending deformation of the structures. The 2.1, set to 12 mm for a twill weave. Therefore, the free-edges aluminium honeycomb located under the front skin is highly boundary conditions chosen for 200-mm side configurations in- deformed in compression under the projectile, with a partial fluencethetargetperformances.Theexperimentaltestconducted densificationatthevicinityofimpact(Fig.11c).Moreover,arupture on200-mmsidetargetislesscriticalthanarealcasebyallowinga of the adhesive bond assembling the honeycomb cells is visible longer yarn sliding distance. Thus, further tests have been con- nearthesupport,showinghighin-planetractionsolicitations.The ductedonbiggersandwichstructuresof300-mmsidetobemore previousobservationsaredirectlyreliedtothetargetdimension,as representativeoftherealcase. for aluminium sandwich structures. Indeed, in 300-mm side
Description: