Influence of strain rate on the deformation and fracture response of a 6061-T6 Al–50 vol Al2O3 continuous-reinforced composite .pdf | Strength Of Materials | Composite Material

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
 4
 
  Materials Science and Engineering A298 (2001) 56 – 62 www.elsevier.com/locate/msea Influence of strain rate on the deformation and fracture response of a 6061-T6 Al–50 vol.% Al2O3 continuous-reinforced composite C.M. Cady *, G.T. Gray III Los Alamos National Laboratory, Mail Stop G755, Los Alamos, NM 87545, USA Received 30 May 2000; received in revised form 12 July 2000 Abstract The compressive mechanical properties of an aluminum – matrix composite unidirectionally reinforced with Al2O3 fibers
Share
Transcript
  Materials Science and Engineering A298 (2001) 56–62 Influence of strain rate on the deformation and fracture responseof a 6061-T6 Al–50 vol.% Al 2 O 3 continuous-reinforced composite C.M. Cady *, G.T. Gray III Los Alamos National Laboratory , Mail Stop G  755  , Los Alamos , NM  87545  , USA Received 30 May 2000; received in revised form 12 July 2000 Abstract The compressive mechanical properties of an aluminum–matrix composite unidirectionally reinforced with Al 2 O 3 fibers havebeen measured and characterized as a function of loading orientation. The influence of strain rate and fiber orientation on thedeformation and fracture response of a 6061 Al–50 vol.% Al 2 O 3 continuous fiber-reinforced metal–matrix composite (MMC)aged to a T6 condition is reported. The stress–strain response of this composite was found to vary substantially as a function of loading orientation; the quasi-static yield changing from nominally 250 MPa transverse to the fibers to  1.7 GPa parallel to thefibers under ideal conditions. Increasing the strain rate to 2000 s − 1 was observed to only slightly increase the yield strength of the composite for both orientations. The main failure mechanism has been identified to be kinking, although an upper boundseems to be attained when the fibers reach their compressive strength. The experimental results are consistent with a plastickinking model for strain hardening composites. The failure response of the composite transverse to the fibers, under both uniaxialstress (quasi-static and dynamic) and uniaxial strain loading, displays a protracted but substantial load drop after yield followedby continued degradation in load carrying capacity. Lack of ideal parallel fiber construction was found to lead to systematicbuckling failure of the alumina fibers through the sample under uniaxial loading. © 2001 Published by Elsevier Science B.V. Keywords :  Aluminum–matrix composite; Loading orientationwww.elsevier.com / locate / msea 1. Introduction Over the last decade, there has been increased indus-trial interest in continuous-fiber reinforced composites,from both the civilian and defense sectors. Polymermatrix composites (PMCs) provide excellent mechani-cal properties such as high specific strength and stiff-ness, making them attractive for use in recreationalsports, automotive, aerospace, and naval structuralcomponents. Metal–matrix composites (MMCs), inparticular continuous-fiber Al–matrix composites (CF-AMCs) are also becoming more widely used. Applica-tions range from landing gear for aircraft, space shuttleand satellite components, to armor for military vehicles.These composites offer exceptional specific propertieswhen compared to monolithic alloys and particulateMMCs [1,2] (Fig. 1). The longitudinal specific strengthsand specific stiffnesses of CF-AMCs are superior tohigh-strength steel and other conventional alloys andtheir transverse properties are generally as good as orbetter than many common steels.Although the mechanical responses of PMCs andMMCs have been extensively studied for various quasi-static loading conditions [3–13], such as tension, shear,and compression, a limited number of high-strain rate,impact, or shock-loading studies have been conductedon fiber reinforced composites. Some studies have re-ported dynamic properties for CF-PMCs [14–16] andparticle reinforced MMCs [14–25]. However, surpris-ingly little research has probed the dynamic response of CF-AMCs [19,22,26]. It is speculated that inelastic mi-crobuckling, fiber crushing and matrix ductility lead tothe failure of CF-AMCs and other CF-composites[3,12–16]. It is also widely known that interfacestrength between the fiber and the matrix plays animportant role in the mechanical properties of CF-AMCs. Tensile properties, in the longitudinal direction,are dominated by those of the fiber and are enhanced * Corresponding author. Tel.: + 1-505-6676369; fax: + 1-505-6678021. E  - mail address :  cady@lanl.gov (C.M. Cady).0921-5093 / 01 / $ - see front matter © 2001 Published by Elsevier Science B.V.PII: S0921-5093(00)01339-3  C  . M  . Cady , G  . T  . Gray III  /  Materials Science and Engineering A 298 (2001) 56–62  57 by weak fiber–matrix interfaces [3,5,6,12,13]. However,transverse and shear mechanical properties are de-graded by a weak interface.In compression, a strong interface and high strengthmatrix are desirable for both axial and transverse com-posite loadings. A strong fiber–matrix interface is desir-able because failure in compression is initially ashear-dominated process. A simple approximation of the strength of an ideal CF-MMC ( | u ) when loadedparallel to the fibers has been modeled as [1]: | u =  fE  f  m  + (1 −  f  ) E  m m  (1)where f  is the fiber volume fraction, E  f  is the Young’smodulus of a fiber bundle, m  is strain at failure, and E  m is the Young’s modulus of the matrix [1]. This rule of mixtures describes the behavior of a CF-compositeassuming ideal and unbroken fibers. Since the matrix isrelatively low in strength compared to the fibers in mostCF-AMCs, it effectively yields immediately and there-fore adds very little to the elastic response of thematerial. This means the composite behavior in CF-AMCs can be approximated solely by the properties of the fibers: | u =  fE  f  m  (2)In engineering materials, a strong matrix is desirableto provide initial constraint in the composite and there-after helps prevent premature fiber buckling. A geomet-rical instability of misaligned fibers is thought to be theinitiation site for failure and the main cause for thelowered compressive strength of CF-AMCs [12,14,15].Much of the understanding of this compressive failuremechanism has been obtained from quasi-static experi-ments, while few studies have been done on the dy-namic compressive behavior of this class of composites.The focus of this study is on an aluminum–alloycomposite reinforced with Nextel™ 610 alumina fibers.This fiber / metal–matrix composite represents a mate-rial which: (1) contains two distinctly different con-stituents in terms of structural, physical, andmechanical properties, (2) exhibits strongly directional Fig. 2. Optical metallography of 3M 6061Al–50 vol.% Al 2 O 3 fiber-re-inforced composite. elastic and plastic anisotropy, and (3) achieves some of its properties due to interfacial effects that influenceplastic flow and fracture behavior. In this paper thequasi-static and dynamic response of a fiber-reinforcedMMC under compressive loading, the failure (fracture)process of the composite, as well as the effect of con-straint on the compressive properties will be presented. 2. Experimental 2  . 1 . Material  A continuous unidirectional fiber pre-form contain-ing  50 vol.% Nextel™ 610 Al 2 O 3 fibers was pressureinfiltrated with a molten 6061 aluminum alloy to pro-duce a porosity-free aluminum–matrix composite (CF-AMC) [27,28]. The AMC, provided by 3MCorporation, was solution heat treated for 2 h at 540°Cthen water quenched and peak aged at 180°C for 6 h.The measured ultrasonic wave speeds normal to thefiber direction for this composite are 7.865 mm m s − 1 for the longitudinal-wave velocity and 4.406 mm m s − 1 for the shear-wave velocity with polarization along thefiber direction. Using the method of cells to representthis composite material, Aboudi [29] obtained analyti-cal expressions for the elastic constants of uniaxial-fibercomposites; the calculated longitudinal- and shear-wavespeeds obtained using Aboudi’s equations are 7.79 and4.02 mm m s − 1 , respectively. The calculated longitudinalwave speed is in very good agreement with measure-ment, but the calculated shear-wave speed is consider-ably lower. The elastic constants used in thesecalculations are 77.37 GPa (0.345) and 252.1 GPa(0.236) for the bulk moduli (Poisson’s ratio) of alu-minum and alumina, respectively [26].The typical fiber distribution in this composite dis-plays non-uniform areas of clustering and small chan-nels devoid of fibers, the latter arising from the initialpacking density of the pre-form and accentuated byfiber movement during casting (Fig. 2). Fig. 1. CF-AMCs offer a good balance of longitudinal and transversespecific strength and stiffness [3].  C  . M  . Cady , G  . T  . Gray III  /  Materials Science and Engineering A 298 (2001) 56–62  58 2  . 2  . Mechanical beha 6 ior The uniaxial-stress mechanical response of the AMCin this study was measured in compression using threedistinct sample geometries. The first is a solid-cylindri-cal sample 5.0 mm in diameter by 5.0 mm in length.The second test sample is 12.0 mm long and 6.3 mm indiameter with a slightly reduced gage section and steelrings 2.0 mm long press fit around the ends of thesample to provide additional constraint (Fig. 3) [14,15].Finite element model calculations were performed toestablish the optimum sample geometry for stress stateequilibrium for the Hopkinson bar samples. This ge-ometry was calculated to produce a uniaxial, uniformstress-state in ceramics when tested at high strain rates[30]. The assumption was made for the CF-AMCs thatthe reinforcing ceramic fibers would dominate the me-chanical properties and the behavior would mimic thatof a ceramic in many key aspects. The final geometryresembles a dumb-bell 12.0 mm long by 6.3 mm indiameter with a reduced section of 5.5 mm length by4.0 mm diameter in the center and a taper from 6.0 to4.0 mm having a 1.0 mm radius. Steel rings were alsoused on these samples to optimize the effect of con-straint (Fig. 3) [14,31]. Geometries 2 and 3 were used tominimize end effects and promote failure within thegage section by suppressing premature axial splitting of the specimen. To some extent this was true. However,when geometry 2 was tested quasi-statically the failurein the sample was observed to be initiated by theslippage of the constraining rings away from the plateninterface and ‘brooming’ was initiated. The constrain-ing rings facilitated a significant increase in the loadthat the fibers supported prior to this slip and subse-quent failure, however the ultimate strength was deter-mined not by the strength of the composite but by thestress at which the ring slipped. Geometry 3 was foundto produce the most accurate and reproducible results.In this case the retaining rings as well as a fraction of the sample diameter act to constrain the ends of thesample. The smaller diameter of the gage section alsodecreased the load necessary to initiate damage andfailure within the sample. Compression tests at strainrates of 0.001 and 0.1 s − 1 were conducted using ascrew-driven load frame. Dynamic tests, strain rates of  Fig. 4. Stress–strain response of the CF-AMC tested transverse to thefiber direction showing one- and two-wave stress curves and strainrate versus strain (all three sample designs). 1000–8000 s − 1 , were performed as a function of strainrate utilizing a Split-Hopkinson Pressure Bar (SHPB)[32–34]. An ASTM specification 6061-T6 Al alloyrolled plate stock material was also tested to facilitatecomparison with the unreinforced matrix Al-alloy. 2  . 3  . Validity of SHPB testing of CF  - AMCs Due to the low fracture strains of ceramics andCF-AMCs loaded axially, the high-rate constitutiveresponse of CF-AMCs needs to be carefully probed toassure valid uniaxial stress SHPB data [32–34].To verify the high-rate SHPB measurements, differ-ent stress-wave analyses [32,33] were calculated to de-termine the specimen stress state from the SHPB barsignals as illustrated in Figs. 4–6. In the one-waveanalysis, the specimen stress is directly proportional tothe bar strain measured in the transmitted bar. Theone-wave stress analysis reflects the conditions at thespecimen-transmitted bar interface and is often referredto as the specimen ‘back stress’. This analysis results in Fig. 5. Stress–strain response of the CF-AMC tested axial to the fiberdirection showing one- and two-wave stress curves and strain rateversus strain.Fig. 3. Specimen geometries: (1) right circular cylinder, (2) geometryafter Cosculluela [30], (3) geometry after Tracy [31].  C  . M  . Cady , G  . T  . Gray III  /  Materials Science and Engineering A 298 (2001) 56–62  59Fig. 6. Stress–strain response of the CF-AMC tested axial to the fiberdirection showing one- and two-wave stress curves and strain rateversus strain for an ‘Overdriven’ sample (geometry 3). 2  . 4  . Damage e 6 olution At low strain rates, samples using test geometry 1and 2 were observed to fail by brooming. There was noconsistent failure stress for either geometry althoughthe constrained sample failed at a substantially higherstress. Geometry 3 produced statistically similar failurestresses for the low strain rate tests with the fractureprocess being identical to that seen in the high strainrate tests.Comparison of the stress–strain behavior of two,high strain rate, axially loaded samples (geometry 3),one of which had just initiated a kink band (Fig. 5) andthe second, which had sufficient energy to fracture andpartially crush the sample (Fig. 6) provides some in-sight into the deformation of these materials. It can beseen that it is possible to ‘overdrive’ the material andcreate a peak stress that cannot ordinarily be achieved.It is clear in both samples that a non-equilibrium stressstate was present in both samples, although the sampletested with a lower pressure, (Fig. 5), appears to becloser to having the requisite characteristics for a validtest (i.e. a constant strain-rate and overlaying one- andtwo-wave stress / strain behaviors). The composite plotsin Figs. 5 and 6 demonstrate the crucial importance of carefully monitoring and utilizing one- and two-wavestress analysis during SHPB testing of this class of material.The strain rate versus strain plot was seen to vary forthe axially loaded samples. A drop in strain rate wasseen to coincide with macroscopic fracture of the speci-men. It is possible for the one- and two-wave stress–strain curves to appear to be in equilibrium (Fig. 6)while the strain rate versus strain plot identifies anon-constant rate. However, there can be non-uniformplastic flow and the slope (or shape) of the strain-rateversus strain plot can change significantly which pro-vides an additional indication of specimen non-uniformdeformation or fracture. Therefore, it is also importantto monitor the strain-rate versus strain for all SHPBtests to validate specimen stress-state equilibrium andto help define the transition from uniform plastic defor-mation to non-homogeneous deformation processes(i.e. such as shear banding) and failure [31].Although many of the assumptions required for avalid SHPB test are not achieved for the current CF-AMCs when tested axially, some conclusions about thematerial behavior can be made with the understandingthat the results are reported for non-equilibrium stress-state conditions. 3. Results and discussion The stress–strain response of the 3M-fiber compositeis seen to vary with fiber orientation and strain rate asmore accurate and smoother stress–strain curves, espe-cially at low strains near yield. Alternatively, the two-wave analysis uses the sum of the synchronized incidentand reflected bar waveforms (opposites in sign) tocalculate the specimen ‘front stress’ which reflects theconditions at the incident / reflected bar–specimen inter-face. A valid, uniaxial stress SHPB test requires that thestress state throughout the specimen achieves equi-librium during the test and this condition can bechecked readily by comparing the one-wave and two-wave stress–strain responses. When the stress state isuniform throughout the specimen, the two-wave stressoscillates about the one-wave stress.Finally, the achievement of a reasonably constantstrain rate is also deemed crucial for obtaining quanti-tative constitutive data suitable for material modelingand as an indicator of deleterious inertial effects ornon-uniform deformation behavior such as shear local-ization or fracture.Based on the observations of the one-wave, two-waveand strain rate analysis during SHPB tests for thecurrent composite, the one- and two-wave data verifiedthat the samples reached an equilibrium stress statewhen tested transverse to the fibers for all three testgeometries. Fig. 4 illustrates that the strain-rate / strainplot is reasonably constant for strain values from 2 to15% for the transversely loaded samples. On the otherhand, based on the criteria listed above for attainmentof a uniaxial stress test, the specimens loaded axial tothe fibers were not found to reach an equilibrium stressstate (Figs. 5 and 6). This does not necessarily totallyeliminate the utility of the |  – m  data on the axial-ori-ented samples, provided that it is well understood thatsome or all of the requirements for valid uniaxial stressSHPB experiments may potentially be compromised.However, data exhibiting protracted stress-state equi-librium should be used with caution and so noted.
Related Search
Similar documents
View more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x