.1___-i"

illl i I1 IJ

illl}

. Low Velocity Impact Fatigue Studieson Glass Epoxy Composite Laminates

with Varied Material and TestParameters - Effect of Incident Energy

and Fibre Volume Fraction

BIJOYSRIKHAN,*R. M. V. G. K. RAo** ANDN. VENKATARAMAN*FRP Pilot Plant

Materials Science DivisionNa1ional Aerospace Laboratories

Bangalore: 560017. India. .d.t,;.:';:t::~;'.

ABSTRAcr,~'~AP effort ~ made to study the effects of fibre volume fcict{'ifn and inci-dent energyo,ll":tl1e impact damage tolerance of composite laminates stioj~ted to lowvelocity impactS':it constant strike velocities. Repeated drop tests were con'd'U'cfb:!using anin~housebuilt':{fu)p'weightimpact-lester: Delamination-area was used as ~.~arameter forquantifying dainage while the number of drops (impacts) to failure use~Fto assess thedamage tolerance limits. The delamination area was found to increase and then saturate

after a certain number of drops. Impact fatigue studies showed the existence of a criticalincident energy (Ec) around which design of cpmposite structures can be based. Also theminimum incident energy required to fracture the sample in a single impact (ESDT)wasevaluated from the data. One of the interesting observations made was that for any givenincident energy, the delamination area was found to be minimum at a certain fibre volume

fraction (0.5 in this case) of the laminate. This was explained on lines of failuremechanisms reported earlier.

..j~'),.I.

INTRODUCTION

IT IS WELL known that fibre reinforced polymer matrix composites undergodamage through a complex process [lt2t3] when impacted by a solid object or

a projectile. This is mainly due to complex interfaces occurring in laminated. composites.Applicationswhere composites are susceptibleto iri1pactby foreignobjects are of particular concern to designers of aerospace structures. Unfor-tunately in many cases consideration of impact resistance of the material is con-spicuously absent in the design either due to the complications arising out of its

"Dcpe of Met2Jlurgy. Kamataka Regional Engineering College, SUJ4thkal. India..oAuthor to whom all com:spondence ~hould be addressed.

1150Journal of REI1':FORCEDPLASTICS AND COMPOSITES, Vol. 14-November 1995

GnI-6844/95/11 1150-10 S10.oo/O@ [995 TcchnDmicPublishingCo., [oe.

-..-.--. . _.-. --. ."-- ~~ ,-"'< ...,>., - '---.

,

Lov.' Velocity Impact Farigue Srudies on Glass Epoxy Corr:posire Laminates 1151

accounting or absence of appropriate information in this field for such materials.Therefore to utilise composites to our full advantage, their response to impactmust be assessed to fairly predictable levels.

Impact induced damage in a composite is a very complex phenomenon consist-ing of a variety of failure modes [1,2,3] that mainly include delamination [4,5],fibre breakage, fibre pullout and total failure of the laminate. The high cost anddifficulty in perfonning the widely discussed [6,7] compression after impact tests

. using sophisticated instrumented drop weight impact testers as a measure of; damage tolerance has necessitated t4e development of a more convenient and less

expensive damage assessment test procedure using an' uninstrumented dropweight impact tester (UDWIT) as adapted in the present studies. Repeated droptesting or impact fatigue is one of the candidate techniques [6,8,9]. Lhymn [8] hasderived a lifetime equation of impact fatigue for PPS/glass composites and ana-lysed it statistically to predict the engineering lifetime for design purposes andalso a minimum impact energy for failure was envisaged. Jang et al. [9] con-ducted repeated low velocity impact tes~ on PPS and epoxy composites using aninstrumented drop weight impact tester and identified threshold incident energyErltand critical impact cycles Nc which were used as indices for ranking of glass,kevlar and graphite fibre reinforced plastics. Wyrick and Adams [6] measured theresidual tensile and compressive strengths of specimens cut from compositeplates subjected to repeated impacts at various incident energy levels.

. -Hbweverreports 'on' repeated~drop' test (RDT)-datausing uninstrumented drop

weight impact testers (UDWIT), which can in a way provide designers with inex-pensive and simple approaches of designing witl1~omposites are not available inopen literature. Also techniques of predicting safe incident impact energy levelsfor composites in service are scant. I f~l(p:

A systematic study involving the effects of bothtmaterial and test parameters onthe impact damage tolerance of laminated cOmJ1rositesusing an in-house builtUDWIT was used by the authors and resulft;~teported. The approach to theproblem should therefore be to study the effectS<ofmaterial parameters like fibrevolume fraction, reinforcement form and its weave style as well as test param-eters on the impact fatigue behavior of composites (see Table 1). The first step inthis direction was to evaluate the effects of fibre volume fraction (VI) and incidentimpact energy (Ein)on the impact fatigue behavior of glass-epoxy laminates andthese results are presented in this paper. Further an attempt was made to com-

Table 1. Factors affecting impact data.

Test Parameters Material Parameters

Incident energyIncident velocityImpactor geometryImpactor materialSpecimen support type

Matrix (brittle/toughened)Reinforcement type (carbon/glass/kevlar/hybrids)Weave styleFibre contentThickness

y

rr

1152 BUOYSRIKHAN, R. M. V. G. K. RAo AND N. VENKATARAMAN

ment on some of the interesting characteristics of impact fatigue curves soobtained. .

EXPERThffiNTAL

Test Laminate Preparation

The test laminates were prepared by a modified compression moulding tech-nique [10]. The matrix used was a room temperature cure epoxy resin. system (LY5052 and HY 5052) supplied by Ciba Giegy (India) Ltd. Reinforcement used wasE-glass satin weave fabric, 8-mil thick. Different number of layers were mouldedto the same cured thickness (2 nun) to yield laminates of three different volumefractions (0.3, 0.5, 0.7). From these laminates test specimens of size 90mm X 90 mm were cut using a diamond edged cutter for repeated drop testing.

Test Procedure

The energy absorption in a composite depen~s ~man array of factofS that im-part specific history to the material under tes~Jl1,12]. It is thus very importantthat data comparisons between various compO.sitesbe made with these factorsclearly specified (which unfortunately is not thec,ase very often) to avoid conflict-ing results. These may be broadly classified into test parameters and materialp¥aIDeters as giv~n.in Table 1. '.

In the present studies the material parameter chosen was the fibre volume frac-tion and test parameter, the incident energy. A stainless steel hemispherical tupof 12.5 rom diameter was used throughout the experiments. The specimen wasclamped along the edges so as to leflvea central circular test area of 11.4sq. mm.All the impacts were carried out at a constant impact velocity of 2.89 mts/sec.The impact tester developed at the FRP Pilot Plant, NAL and used in the studiesis shown in Figure 1.

The test procedure consisted of repeatedly impacting the clamped specimen tillpenetration occurred. Delamination area gIOwthand number of drops to failure(Nf) were chosen as damage tolerance assessment parameters and noted for dif-ferent Ei" values and for specimens of three different Vf values. The delaminationarea (whitened region created upon impact) was traced to a graph sheet and areasmeasured as a function of the drop number.

RESULTS AND DISCUSSIONS

The data are plotted as Ejn vs Nf and delamination area vs drop number for dif-ferent Vf values and delamination area vs Vj for different Ein values. The observedresults are discussed in the following paragraphs.

Effect of Ebt on Nf

Figure 2 shows the plot of Ejn vs Nf for specimens of three different V/s repre-senting resin rich and fibre rich composites. It can be seen that all the curves (theimpact fatigue curves) exhibit a characteristic knee point (Figure 2), correspond-

.---------..-. . -. """""---:-~,,,,,,,':;:'.:::'-'---: - .,---" .

Low Vel~city Impacr Fatigue STudies on Glass Epoxy ComposiTe LaminaTes

r'

;

-->:;'1'1" ",'--

">' J~'~b-rl~~

"'~"r,>' 1'

,

-,1'., ,r

~Ufi;,'j-td

0

II

IJ

iV'( 8 I0

1. :;u L!~ ::-.1:0., 2. ::le<:tro :naqnet 3. :;;r-:;::-.." ight ... Rope dr_H;'

s. ;'~r.anlcalcounter 6. 5pecd reducer ::JO-dr'::>ox 7. :::;lectrlc ~t:::r

9. :,c su;::p!:r '.1:-:it 10. ?!:-:-::--'re !.-;,C' ~ol~~nq 3peci"""l1? '>41re rope

11. =u?

Figure 1. The NAL drop weight impact tester.

1153

.;,

-r

/

'"

"

1154 BUOYSRlKHAN, R. M. V G. K. RAo AND N. VENKATARAMANj

0

15 -+-

10

0

Vf. 0.3+ 0.50 0.7

5

0 I ,.,

1 It 41Figure 2. Impact fatigue curves f?'fi}¥J.i11inatesof three different volume fractions. X-axjs:number of drops to failure N,; Y-axfPiiilJcident energy Einjoules.

;.'"f.'r\':

ing.tO-which a..criticaLincidenU;;~:\1J£igy.£;,.canbe identified as that value for which!1Ei../ ~j - 1. It is clear that',Nf decreases drastically for values of Ei" > Eo

signifying damage susceptibility and the vice versa. A design incident energylimit Ed can therefore be evolved such that (EdIE,,) < 1. The choice of the ratio(EdlEe) has to be chosen discretely ,by the designer, considering the probabilityof impacts in service. Higher the probability of impacts, lower should be (EdIEJ.If the expected Ei.. > Ee. then an optimization among various material parame-ters (Vj, thickness, matrix toughness, fibre nature, etc.) has to be struck in thedesign process.

Further, if tests are conducted at lower Ej", an incident energy level may beidentified below which NI tends to infinity. This incident energy level is thus anal-ogous to endurance limit as found in S-N diagrams used for fatigue analysis.However due to practical considerations in conducting so many impacts, such arattempt was not made. .

Analysing the plots shown in Figure 2, it can be possible to establish a correIa.tion between Ei.. and NI through a theoretical relationship. From the data obtaineiit is seen that Ei.. varies inversely as the power of Nj,

Ein (X l!N}

E'n = A(Njtb

where A is the constant of proportionality and serves as a parameter for quantifying damage tolerance. This can be better understood when we substitute Nf = 1Then E,.. = A, which is nothing but the minimum energy required to fracture thl

___m ...... : '-'''-'C'-::C.-'~..''., ~-~- - .- "":",-.C' .~---~

Low Velocity Impact Fatigue Studies on G/pss Epoxy Composite Laminates 1155I

;I'It

0

10

>-

11 10

'.'~ . -.- -",Fig~re 3. Log-Logplot of impact fatigJ,!edata for laminates of three different volume frac-

,'" ,... "'Hans:X.axis: niJiilbetof drop~;tO'fat1iir:'lf'N,;'Y-fixis:incident energy Ei~joules.

>~ .>"'"

---1.11'.. .-nu.

specimen in a single imR{tCfk(S.D.T'ivalue) at the given incident energy. From tiie~:<log-log plot (Figure 3h~,:;see that the slope (b) is more or less constant fou!'!rxgiven set of material p3.!fWleters.For the range of volume fractions investigate:.d\!~1b varied between -O.37land -0.41. The values of constants A and b for diffenhuoc1',-"--. . .VI values are as given iit~~:rable2. . FP-

Now referring to Figuiis 2 and 3 it may be observed that Ecand A increase withVI signifying an impro\'ement in impact damage tolerance. Table 3 gives Ecvalues for different Vis. It is worthwhile to point out here that these values arecharacteristic of the particular material, support conditions and impact velocity.

Effect of Drop Number on Delamination Area

In order to correlate between delamination mechanism and energy absorption

Table 2. Values of constants.

V, A b

0.3 14.13 -0.370.5 17.32 -00400.7 19.85 -0.39

--- ----

1156 BUOYSRIKHAN, R. M. V. G. K. RAo ANDN. VENKATARAMAN

Table 3. Critical energy values fordifferent "V,'s".

"

V, Ec.-

0.30.50.7

9.0259.812

11.173

process in RDTs, the delamination areas were measured at the end of each impactand then plotted against drop number for different Vfvalues as shown in Figure 4.

It can be noted that the delamination area grows initially as the number ofimpacts increases and reaches a saturation limit. The initial linear increase indelamination area and its attaining a saturation value, clearly implies that theenergy absorption (and obviously a major part of the rotal energy absorbed) is byand large due to delamination process to start with. Once the saturation isreached further impacts are absorbed by other failure processes likedibre 'break-age and pull~m\Jeading to total penetration of the tup through the 1a:1(1lpate.Thisresult once ag~in confirms the reasons as to why delamination rend~Cy of lami-nated compqsj,tes is by far regarded as the most critical composite :'GhaFacteristic,., "J..< - -,-

that decides:'th~IT impact damage resistance. The delamination arcth.growths as'- :tTatetfotnr trapsparent'sheet-<are'presented'in -Figure 5 for different:£;11values.

The decrease:ih the number of drops to failure with increase in E... is dearly seenfrom this schematic figure. .

Effect of VI on Extent of Delamination

From Figure 6, it can be observed that the delamination area first decreases andthen increases as VIis increased. The same trend is seen to prevail for all incident

300-

200ooooooooooooooooN f00

000

00

00

0

. 100- 0

0

0

0

0

00I

10l

30I

1.0'20

Figure 4. Delaminationarea vs drop number. Y xis:drop number N; Y-axis: delaminationarea mml.

,--~ ,.. . ~ -.- '~---

T

-...--.

Low J-elociry Impacr Farigue Studies on Glass Epoxy Composire LaminaTes

01

1157

0 0000~

2 s 63 .

AT IlX:~:Jc:::r i:liERGY OF 7.10 j au1..

01.

0 00AX IN:ID£Nr Et«RGY or 9.89 jouua

2 3 4

Rgure 5. Delamination area growths as traced on a transparent sheet for laminate ofvolume fraction0.5 at two Einlevels.

energy levels and at any drop number, with the delamination area reaching aminima at a typical VI = 0.5. This is a very interesting feature observed in all thecomposite laminates tested. This phenomenon can be attributed to the two failure

. m,~h~is.m&.{1;8.J3l,dominatingH-in,;diff'efeflt.,-VI'domains"as clearly shown inFigure 7. The two mechanisms are as follows

. Extensive matrix crackings percolatihg down to interface region leading todelamination (dominant below VI .:- 0.5).

. Decreasing ILSS values causin~~~'~;oss delaminations (dominant aboveVI = 0.5). .. .. ,. . .

} uO'.

Thus a minimum delamination is mpJEcedat the cross over point of these twomechanisms. ..

200

100

0.2 0.3 0.L. 0.5 0.70.6

Figure 6. Delamination area vs fiber volume fraction at three incident energy levels. X-axis:fioor volume fraction Vj; Y-axis: delamination area mm2.

II"

11.)~ BUOYSRlKHAN, R. M. V. G. K. RAo AND N. VENKATARAMAN

'" /'

~",/

'","/'""

, "De 1 'i1II1natlon "x

."./''''governed by ,,"znat.rix craclcinq .".

/' ...

""" ~,. " OPr:IK1M '

--""", ::>ELUUNo\T:ION '-----

~..laminat1oD7ov~ by:u.ss

0 0-&

Figure 7. Damage mechanisms operative in different V, domains. X-axis: Fiber volume; fraction V,; Y-axis:Delaminationarea.

CONCLUSIONS ".~: .'.

An uninstrumented drop weight impact tester. has been built and effectively. used for impact damage tolerance characterisation of composites laminates.; As the incident energy of the impactor is increased the number of drops to fail-, ure decreases. There exists a critical incident em~rgx value which can be evalu-t ated from ~the'-impact'fatigi1(;n:llrves. '. .'"C-'"" . -. ".' .'" .

'. Delamination area grows with energy absorption and reaches a saturation value: indicating that energy absorption process changes over from one of delamination; to that caused by fibre breakage and pullouts which ultimately lead to penetration.of laminate. '.

: Increasing the fibre volume fraction in general results in improved damage-tolerance., The extent of delamination was observed to be least at around VI = 0.5, whichsuggest that an optimum delamination occurs when a transition in energy absorp-tion mechanisms take place.. .: Further studieson other materialand testparameters(in progress)are expectedto give a more comprehen~ive picture of the impact damage tolerance in lami-Gatedpolymer composites.

ACKNOWLEDGMENTS

! The authors are thankful to Dr. K. N. Raju, Director, Prof. R. Narasimha,onner Director and Dr. A. K. Singh, Head' Materials Science Division, Na-ional Aerospace Laboratories for all the support they received in the investiga-ions. They further acknowledge the invaluable assistance received from Mr. K.:uresh Raju in the preparation of this paper. .

NOl\1ENCLATURE

Ej" = incident energy

. - " .-:.; ;.: . . - ~ ';-:. ....- - '- .'---

" :.~.

rr

'.

Low ~lociry Impact Farigue Studies on Gl£us Epoxy Composite Laminates 1159

Ec = critical incident energyEd = design incident energyV, = fibre volume fractionN = drop/impact numberN, = failuredropnumber

S.D.T. = single drop test

REFERENCES

1. Cristescu, N., L. E. Malvem and R. L. Sierakowski. 1975."Failure Mechanisms in Compos-ite Plates Impacted by Blunt Penetrators," Foreign Object Impact Damage to Composites, ASTMSIP 568, American Society for Testing and Materials, pp. 159-172.

2. Agarwal, B. D. and L. 1. Broutman. 1980. Analysis and Performance of Fibre Composites. John, Wiley and Sons.

3. Broutrnan, L. J. and A. Rotem. 1975. "Impact Strength and Toughness of Fiber Composi~Materials; Foreign Object Impact Damage to Composites, ASTM STP 568, American Society forTesting and Materials, pp. 114-133. '.

4. Amar, C. Garg. 1988. "Delamination-A Damage Mode in Composite Structures," Engineering

Fracture Mechanf.cs, 29:557-584.5. Liu, D. 1988. "Impact Induced Delamination-A View of Bending Stiffness Mismatching," J. of

Composite Materials, 22:674-692.

6. Wyrick, D. A. and D. F. Adams. 1988. "Residual Strength of CarbonJEpoxy Composite Material'S'uDjeeted' to 'Repeated"Impact;~ J:"oFCinnpusite 'Materio.Is~"2-2(8):749~765.'

7. Prichard, 1. C. and P. J. Hogg. 1990. "The Role of Impact Damage in Post-Impact Compression, : Testing," Composiles. 21(6):503-511. , '

!;~>'8. Chang, L. 1988. "Impact Fatigue of PPSIGlass Composites,~J. of Materials Science Leuers,

4:1221-1224. i' ,

.;Ek,Z9. lang, B. P., C. T. Huang, C. Y. Hsieh, W. Kowbel and:B,:: Z. lang. 1991. "Repeated Impact"z,rxii, Failure of Continuous Fibre Reinforced Thennoplastic'and, Thermoset Composites," J. of

, :. " Composite Materials. 25:1171-1203. . .;; .,'" ."~:';,191O. Kaushal, S., K. Tankala, R. M. V. G. K. Rao and Kishore/:1991. "Some Hygrothermal Effects

"",".J~~i'- . on the Mechanical Behavior and Fractography of Glass-Epo~Y'Composites with Modified Inter-;<' face," J. of Materials Science, 26:6293-6299. J{

11. Cantwell, W. J. and J. Morton. 1991. "The Impact Resistance of Composite Materials-A Review," Composites, 22:347-362.

12. Cantwell, W. 1. and J. Morton. 1989. "Geometrical Effects on the Low Velocity Impact Responseof CFRP," Composite Strucrures, 12:46-53.

13. Khan, B. 1992. "Impact Fatigue Behaviour of and Damage Mechanisms in Glass-EpoxyComposite Laminates," M. tech thesis, KREC/NAL.

",