HAL Id: jpa-00224782
https://hal.archives-ouvertes.fr/jpa-00224782
Submitted on 1 Jan 1985
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
EFFECT OF STRAIN RATE ON THE TENSILE FAILURE OF WOVEN REINFORCED POLYESTER
RESIN COMPOSITES
L. Welsh, J. Harding
To cite this version:
L. Welsh, J. Harding. EFFECT OF STRAIN RATE ON THE TENSILE FAILURE OF WOVEN REINFORCED POLYESTER RESIN COMPOSITES. Journal de Physique Colloques, 1985, 46 (C5), pp.C5-405-C5-414. �10.1051/jphyscol:1985551�. �jpa-00224782�
JOURNAL DE PHYSIQUE
Colloque C5, supplCment au n08, Tome 46, aoQt 1985 page C5-405
EFFECT OF S T R A I N RATE ON THE T E N S I L E F A I L U R E OF WOVEN REINFORCED POLYESTER R E S I N COMPOSITES
L.M. Welsh and J. Harding
Department o f Engineering Science, U n i v e r s i t y o f Oxford, Parks Road, Oxford O X 1 3PJ, U.K.
Resume - Des courbes contrainte-dbformation determinbes en t r a c t i o n b des v i t e s s e s de %lo-&, 10 e t 1000s-I s u r l e s composites renforces par des t i s s u s de carbone, Kevlar ou v e r r e sont presentees. On determine l e s v a r i a t i o n s du module d e l a c o n t r a i n t e e t de l a deformation b rupture en f o n c t i o n de l a v i t e s s e de dbformation. A l ' i n v e r s e du comportement des composites carbone e t Kevlar, on observe dans l e c a s des composites renforces p a r f i b r e de v e r r e un changement de mode de rupture 1 grande v i t e s s e de d8formation.
Abstract - Tensile stress-strain curves a r e obtained at s t r a i n r a t e s of %lo-'+, ~ 1 0 and c1000/s for carbon, Kevlar and glass fibre woven-reinforced polyester r e s i n compo- s i t e s and the variation of modulus, tensile strength and f r a c t u r e strain with strain r a t e is determined. A change in the failure mode at the highest strain r a t e s i s observed for the glass but not the carbon o r Kevlar reinforced composites.
I - INTRODUCTION
The increasing use of fibre-reinforced plastics in situations involving dynamic loading, their developing application in energy absorbing systems associated with impact and their suscepti- bility to micro-damage under sub-critical impact, leading to a subsequent deterioration in mechanical properties a t conventional Loading rates, has resulted in an urgent need for funda- mental studies of the effect of the r a t e of loading both on the mechanical strength and on the mechanisms of energy absorbtion in these materials. Unfortunately much of the previous work in this area has used Charpy o r Izod type impact bend tests where stress-wave reflections and the complex modes of failure involved inhibit any fundamental analysis of the material response and of the effect of Loading r a t e on the fracture hehaviour. Although an instrumented tup may be used and load-time records obtained, allowing estimates of both the energy absorbed a t the initiation of fracture and of the total work of fracture (1,Z) to be made, the s t r e s s system under which failure occurs i s complex and a fundamental study of material response i s not possible.
Other types of test which have been used include those employing a drop weight (3), where s t r e s s oscillations on the load cell signal obscured the details of the process and limited the maximum loading rate, and those employing a pendulum (4) or fIywhee1 (5) to apply a direct tensile impact.
In both these latter methods doubt remains regarding the accuracy of the strain measuring tech- nique, which i s particularly important when testing materials with such relatively low strains to failure.
In the second of these, the method developed by Kawata e t al. (5), a version of the split Hopkinson's P r e s s u r e bar is employed. This technique, although widely used for the impact testing of
metallic specimens, presents some difficulties when applied to composite materials, particularly a s regards the design of the specimen. Recommended specimen designs for the uniaxial quasi- static testing of composite materials (6) a r e not suitable f o r use in the Hopkinson-bar (7). This problem has, to some extent, been avoided in tests using shear loading configurations (8, 9) but the greatest difficulty has been experienced in obtaining reliable data for impact tension. However,
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1985551
C5-406 JOURNAL DE PHYSIQUE
in a recent paper (10) a newly developed technique for the tensile impact testing of fibre- reinforced composites, based on a modified version of the standard tensile split Hopkinson's p r e s s u r e bar (SHPB), has been described. Results w e r e presented for unidirectionally- reinforced CFRP, for which no effect of tensile strain r a t e was observed on the modulus, failure strength or fracture appearance over some seven orders of magnitude, and, in a subsequent paper (I]), f o r woven-reinforced GFRP, where the modulus and failure strength both increased with strain r a t e and there was a marked change in the fracture appearance. Examination of the broken specimens showed considerable matrix damage preceding final fracture in the GFRP tests but not in those on CFRP specimens. A tentative phenomenological description of failure in woven GFRP specimens was proposed. This related (i) the r a t e dependence of the modulus to the effect of the epoxy r e s i n matrix resisting the straightening of the axially-aligned rovings undel increasing load, (ii) the subsequent reduced stiffness ('knee effect') to cracking in the matrix and a breakdown of the bonding between the matrix and the reinforcement and (iii) the r a t e de- pendence of the fracture strength and failure strain to the increased strength of glass f i b r e s ,at impact rates allowing a greater degree of straightening before final failure. In contrast, the strain-rate independent behaviour of the CFRP specimens was related to the unidirectional nature of the reinforcement, where the straightening mechanism no longer applies, and an apparent r a t e insensitivity of the fracture strength of carbon fibres. In these preliminary tests, however, a direct comparison between the impact response of carbon and glass-fibre reinforced composites was not possible because of the different reinforcement geometries in the two com- posites studied.
Nevertheless, it follows from the above interpretation of these preliminary r e s u l t s that a CFRP specimen with a woven-reinforcement geometry might show both a r a t e dependent tensile modu- lus and an increasingly non-tinear stress-strain response a t high rates, provided the resistance of the matrix to the straightening of the carbon fibres contributes significantly to the total load carried by the specimen. However, a r a t e dependence of the fracture strength, if it depends only on the fibre properties, would not be expected. To investigate these possibilities, tests have been performed a t tensile strain r a t e s of %O. O O O l / s , %lO/s and +700/s on CFRP specimens with a satin-weave reinforcement geometry in a polyester r e s i n matrix. To allow a direct com- parison of the effect of the type of reinforcing fibre on the strain-rate senstiivity of fibre- reinforced plastics, s i m i l a r tests have also been performed on KFRP and GFRP specimens with the s a m e satin-weave reinforcement geometry and the same polyester resin matrix. The effect of strain r a t e on the tensile modulus, fracture strength and fracture strain is determined for the three types of specimen and the modes of failure a r e examined.
11' - EXPERLMENTAL DETAILS
Since the impact testing technique has been fully reported elsewhere (11) only a brief description will be given here. Thin strip specimens, waisted in the thickness direction, a s shown in fig. 1, a r e fixed with epoxy adhesive in parallel-sided slots in the yoke and the inertia bar, s e e fig. 2.
Strain gauges mounted on the inertia bar monitor both the s t r e s s transmitted by the specimen and the velocity at the upper end. A calibration for the velocity a t the lower end of the specimen is obtained f o r a given impact velocity by the technique described previously (11) and the strain determined using the SHPB analysis. Signals from the inertia bar and specimen strain gauges were stored in a Datalabs type 922 transient recorder and subsequently displayed on an oscillo- scope, hard copy being obtained on an X-t chart recorder. Typical oscilloscope traces for the inertia bar strain gauge signals in impact tests on the three types of composite a r e shown in fig. 3, the total sweep time being 100 us. Comparative t e s t s at low and intermediate r a t e s w e r e ob- tained on the same design of specimen using a standard screw-driven Instron loading machine, for a strain r a t e of 'LO. 0001/s, and a hydraulically-operated Loading machine, f o r a strain r a t e of %10/s. In at least one test a t each strain r a t e on each type of specimen a check of the strain in the elastic region, and hence a confirmation of the specimen modulus, was made using strain gauges attached directly to the specimen. Typical traces, a s displayed subsequently on an
oscilloscope screen, for a test on a GFRP specimen in the hydraulic machine a r e shown in fig. 4.
' I
, 1 1 1 1 11 . 111111 1'1 1. 111 11 111 1I I I
1 1 . 1I I I ~ I I I I U I I I IIIIUIIIOII IIUIIIII 0
- 1 1 1 , q 1-1 1 1 e ~ 1 1 1 ~ 11
-
1 1 -3 ~-,,,- -. .-,.,- -...-
Fig.1 SPECIMEN DESIGN AND REINFORCEMENT CONFIGURATION ( a l l dimensions i n m i l l i m e t r e s ) Yoke
Fig.2 TENSILE SHPB
Weighbar tube
Inertla bar
Strain gauges
Specimen
Lower grip
a ) CFRP b) KFRP c ) GFRP
Fig.3 INERTIA-BAR STRAIN GAUGE SIGNALS FOR IMPACT TESTS ( t o t a l sweep time - 100ps)
The total sweep time was 4 milliseconds. The upper trace was derived from the s t r a i n gauges attached directly to the parallel region of the specimen gauge length.
I11 - SPECIMEN MATERIALS
Specimens w e r e cut from 3mm thick Laminates supplied by Fothergill and Harvey Ltd. Each laminate consisted of 8 reinforcement Layers p r e s s moulded in Crystic 272 polyester resin to give a calculated volume fraction of 40%. A 5-end satin weave construction was used for each fabric with 6.7 tows/cm in both warp and weft directions f o r the Kevlar and the glass fabrics and 7.0 tows/cm for the carbon fabric, with corresponding fabric weights of 219.4, 353 and 283 g / m 2 respectively. The fibres were, respectively, kevlar 49 (type 968), E-glass (EC13) and Toray type T300-3000 filament carbon tows. The reinforcement geometry of the three types of speci- men is shown, to the same scale a s the specimen, in fig. 1. Within the reduced specimen gauge thickness there a r e 4 layers of fabric while across the full width of the specimen there a r e con- tained between 6 and 7 longitudinal tows. This relatively coarse reinforcement geometry, in relation to the specimen dimensions, was dictated by the use of 3000 filament carbon tows in the woven carbon fabric due to the difficulty of obtaining carbon fabrics woven from the finer 1000 filament tows. It compares with the plain weave GFRP composite tested in the e a r l i e r work (11) where, for similar specimen dimensions, there were 9 layers of fabric in the specimen gauge thickness direction and 18 tows across the full width. In consequence an increased experimental s c a t t e r may be expected in the present tests. Although this could have been Limited by in- creasing the specimen dimensions in proportion to the scale of the reinforcement geometry a corresponding increase in the stress-wave transit time across the specimen would introduce e r r o r s into the Hopkinson bar analysis in the elastic region.
JOURNAL DE PHYSIQUE
IV - RESULTS Stress-Strain Curves
Stress-strain curves for the CFRP specimens a r e presented in fig. 5 for 4 tests a t each of 3 strain rates, nominally 0.0001/s, 10/s and 700 /s. In all tests some inelastic deformation of the speci- men was observed beyond the region of approximately linear stress-strain response. At the quasi-static loading r a t e audible cracking of the specimen was noticed a t around the Limit of the linear elastic region. Wide differences in the s t r e s s Levels associated with the subsequent in- elastic deformation a r e apparent when the 4 specimens tested at any given strain r a t e a r e com- pared. As discussed in the previous section this behaviour i s thought to represent genuine dif- ferences in response between the various specimens due to the relatively coarse nature of the woven reinforcement geometry in relation to the overall specimen dimensions. It i s not possible, therefore, to describe the mechanical behaviour of the composite in t e r m s of a mean, o r average, stress-strain curve a t a given s t r a i n rate. Instead, the results have been characterised by determining, for each test, the initial slope (or tensile elastic modulus, E), the maximum s t r e s s supported by the specimen, (or tensile strength, IS,), and the strain to fracture, E f , defined a s the strain corresponding to the maximum s t r e s s . Despite the wide scatter it is clear that there i s a significant increase in the maximum s t r e s s supported by the specimens at the impact load- ing r a t e and that this i s accompanied by an increase in the Linear elastic range. A wide scatter i s also apparent in the strain to f r a c t u r e at each s t r a i n r a t e such that the apparent slight increase in fracture strain a t intermediate strain r a t e s i s probably not significant. Taking all the CFRP tests together, however, the average fracture strain, 1.59%, is significantly greater than that for unidirectionally reinforced specimens, %0.9%, determined in the e a r l i e r investigation (10).
Stress-strain curves for the KFRP specimens a r e presented in fig. 6 a t the s a m e three nominal strain rates. The behaviour i s very similar to that shown by the CFRP specimens with again considerable scatter in the s t r e s s Levels corresponding to the region of inelastic deformation but, despite this, again a clearly marked increase in the maximum s t r e s s at the impact Loading rate.
At all testing rates the failure strains for KFRP specimens exceeded those for the CFRP speci- mens but with a trend in the KFRP specimens f o r the failure strain to decrease slightly, but con- tinuously, with increasing strain rate.
A markedly different behaviour i s found, however, when the stress-strain curves for GFRP speci- mens, presented in fig. 7, a r e examined. These show a very significant increase in both the maximum s t r e s s and the fracture strain between the quasi-static and the intermediate r a t e tests.
At the intermediate strain r a t e the stress-strain curve shows a very pronounced 'knee effectt with a much reduced stiffness at strains from %20/0, the limit of the linear elastic region, up to failure at -7%. However, on increasing the strain r a t e further, to just over 1000/s, a change in behaviour i s observed. The maximum s t r e s s i s now reduced and corresponds to the s t r e s s level a t the limit of the linear elastic region. Subsequent deformation is at an approximately constant o r slightly decreasing Load to a sufficiently high s t r a i n that failure was not reached in the standard sweep time of loops, see fig. 3c. In one impact test on a G F R P specimen the sweep time was increased to 2 0 0 ~ s . Although this reduced the accuracy of calculation in the low strain elastic region it did allow the entire loading history to be recorded. In this test the f r a c t u r e strain was calculated to be -13%
F r a c t u r e Appearance
Despite the apparently increased extent of inelastic deformation shown by CFRP specimens at the intermediate loading r a t e no effect of s t r a i n r a t e on the fracture appearance could be detected.
At all strain rates specimens broke in the central region of the gauge section, s e e fig. 8a. This figure also shows how the fracture surface i s related to the reinforcement geometry. The in- elastic deformation a t high loads i s clearly related to the long pull-out lengths of the fibre tows, more easily seen when the specimen of fig. 8a i s viewed in silhouette, a s in fig. 8b. This be- haviour may be compared with that found in the e a r l i e r work on unidirectionally-reinforced CFRP (Hyfil) where again the fracture appearance was unaffected by strain r a t e but the pull-out lengths
GFRP Test: sweep time, 4ms a) Specimen strain
b) Specimen load
c) Crosshead displacement Fig.5 CFRP STRESS-STRAIN CURVES
600
Stress-(MPa)
LOO
300
200
100 /
KFRP STRESS-STRAIN
,'.., , --- .
, \ , , ,' , . 7 - > ~ ~ ~ ~ . .-.- y:Z'\x.
1260h
Strain-(per centl
GFRP STRESS-STRAIN CURVES
-
a) CFRP b) CFRP c) KFRP
(reflected light: x3) (transmitted light: x3) (transmitted light: x3) Fig.8 FRACTURE APPEARANCE OF CFRP AND KFRP SPECIMENS
C5-410 JOURNAL DE PHYSIQUE
w e r e v e r y s h o r t and the s t r e s s - s t r a i n r e s p o n s e w a s n e a r to linear-elastic through to failure.
Examination of broken K F R P specimens a l s o showed the f r a c t u r e appearance to be r a t e indepen- dent. Although f r a c t u r e s t r a i n s w e r e marginally higher than f o r CFRP, pull-out lengths w e r e a little s h o r t e r , s e e fig. 8c. However, this may b e a somewhat misleading observation since, although the pulled-out f i b r e tows of C F R P (and G F R P ) w e r e stiff and brittle, those of K F R P w e r e soft and flexible and t h e r e w a s s o m e indication that the polyester r e s i n had not wetted-out the Kevlar f i b r e s very well s o that the m e a s u r e d mechanical p r o p e r t i e s f o r the K F R P specimens reflected m o r e closely the behaviour of d r y Kevlar fibres.
As in the e a r l i e r work ( l l ) , and unlike the behaviour of C F R P and K F R P described above, a m a r k e d effect of s t r a i n r a t e on the f r a c t u r e appearance of GFRP specimens w a s observed. At the quasi-static r a t e , s e e fig. 9a, the m a i n tensile f r a c t u r e w a s a t one end of the p a r a l l e l gauge section with s e v e r a l pulled-out f i b r e tows extending into the tapered region but not a s f a r a s the g r i p region of the specimen, i. e. not within the parallel-sided slot of the loading-bar. F i b r e tows will have also pulled-out f r o m the half of the specimen shown in fig. 9a. At the intermediate loading r a t e the main tensile f r a c t u r e again occurred, in the c a s e shown in fig. 9b, just within the specimen parallel gauge region, but t h e r e w e r e both a g r e a t e r number of f i b r e tows pulling out and a markedly increased a v e r a g e pull-out Length. The longest of the pulled-out f i b r e tows ex- tended throughout the whole of the p a r a l l e l gauge section, the whole of the tapered section and a f u r t h e r few m i l l i m e t r e s into the region within the loading b a r slot. On f u r t h e r increasing the s t r a i n r a t e to %1000/s, in the impact machine, a change in f r a c t u r e mode w a s observed, s e e fig.
9c. T h e specimen disintegrated into t h r e e pieces a s the c e n t r a l portion remained intact but pulled-out f r o m the two ends which w e r e s t i l l attached to the loading bars. When these end- pieces w e r e c u t off f r o m the loading b a r s they e a c h separated into two halves r e p r e s e n t i n g the tapered regions e i t h e r s i d e of the c e n t r a l pulled-out region. The longest pulled-out f i b r e tows extended a t one end a s f a r a s the end s u r f a c e of the original specimen. All other Longitudinal tows failed in tension at s o m e point in the tapered o r g r i p regions. No c l e a r tensile f r a c t u r e s u r f a c e s w e r e apparent a t either end of the c e n t r a l region which w a s longer viewed f r o m one side, 18.8mm o r about 12 f i b r e tows, than viewed f r o m the other, 6 m m o r about 4 f i b r e tows.
V - DISCUSSION
Effect of Strain Rate on the T e n s i l e (Youngs) Modulus
The variation of Youngs modulus with s t r a i n r a t e f o r the t h r e e composites i s shown in fig. 10.
The experimental a c c u r a c y with which the modulus c a n be m e a s u r e d in the tensile testing tech- niques used h e r e h a s been discussed previously (11) and i s expected to be about ?50/c. The much g r e a t e r s c a t t e r here, of the o r d e r of +15% in the C F R P tests, i s thought to be r e l a t e d to the c o a r s e n e s s of the reinforcement geometry, a s suggested in section ID. Nevertheless, despite the s c a t t e r in the r e s u l t s of the C F R P t e s t s a r e m a r k a b l y c l o s e a g r e e m e n t i s apparent in the s t r a i n r a t e dependence of the tensile modulus f o r the t h r e e m a t e r i a l s . Since the reinforcement geometry and the m a t r i x r e s i n a r e the s a m e f o r a l l t h r e e composites, this gives s t r o n g support to the e a r l i e r proposal that the .ate dependence of the modulus in woven-reinforced composites d e r i v e s f r o m the e l a s t i c interaction between the axially-aligned f i b r e tows and the r e s i n m a t r i x and depends on the rate-sensitivity of the matrix. T h e only difference between the t h r e e com- posites tested h e r e i s in the type of reinforcing fibre. This leads to the different absolute values of tensile modulus a t any given s t r a i n r a t e such that E <Ep( <EC w h e r e suffixes G, K and C denote GFRP, K F R P and CFRP respectively. G
change in the rate-dependence of the modulus might be expected f o r different m a t r i x r e s i n s o r reinforcement geometries. Thus, for GFRP, a g r e a t e r i n c r e a s e in modulus with s t r a i n r a t e w a s apparent in the e a r l i e r t e s t s on a fine plain-weave epoxy r e s i n composite, s e e f i g . l l , than w a s found in the p r e s e n t t e s t s on a c o a r s e satin-weave polyester r e s i n material. T h i s i s in accord with the g r e a t e r e l a s t i c interaction that would be expected between the m a t r i x r e s i n and a fine plain-weave than a c o a r s e satin-weave reinforcement geometry. Since the f i b r e volume f r a c t i o n s f o r the two types of GFRP w e r e v e r y s i m i l a r , the difference in absolute modulus values in fig.11 i s presumably related t o the different p r o p e r t i e s of t h e two m a t r i x r e s i n s .
Parallel re~io_n_ - - - - -
--- --- -l
I
+-- -,,--- -,-- --- 4
a) Quasi-static test (x2.1)
Fracture surface
-
---.. Parallel region ---,- -- --- 1 I,,,,,,,,,,-,-,,J I
b) Intermediate rate test (x2.1)
I *
c) Impact test (XI
a L
-
C
Fig.9 E F F E d S T R A I N WITON FRACTLIRE AF'PEARANCE OF GFRP SPECIMENS
C5-412 JOURNAL DE PHYSIQUE
Effect of Strain Rate on the Tensile Strength
The variation of the tensile strength with strain r a t e for the CFRP and KFRP specimens i s shown in fig. 12. Again the experimental scatter i s quite Large. Nevertheless both composites clearly show very similar behaviour with a nearly rate-independent strength of %400MPa up to a strain r a t e of d O / s and then a small, but significant, increase in strength, by about 25%. a s the strain r a t e i s raised to the o r d e r of 1000/s, the KFRP having a marginally lower strength than the CFRP.
Previously it was assumed that the fracture strength of carbon fibres was r a t e independent. If the present increase in tensile strength at impact strain r a t e s i s not due to an increased strength of the carbon (and the Kevlar) fibres it suggests that the resin itself c a r r i e s a significant proportion of the total load in woven reinforced specimens at the highest s t r a i n rates. This could be the case, provided the bonding between the resin and the reinforcement remains largely intact s o that the elastic interaction between them is still possible. The observation that the fracture strains for CFRP and KFRP a r e relatively insensitive to strain r a t e and perhaps decrease slightly a t the highest r a t e s Lends support to this proposition.
In previous work on woven GFRP ( l l ) , s e e fig. 13, a much g r e a t e r rate-sensitivity of the tensile strength was observed, an increase of %70% being recorded between the quasi-static and inter- mediate r a t e s of s t r a i n and an overall increase of ~ 1 6 0 % between the quasi-static and the impact r a t e s of strain. The present results also show a significant r a t e sensitivity between and 10/s, the tensile strength increasing by%50%. In the impact tests, however, this behaviour i s reversed, the tensile strength showing a decrease of ~ 2 5 % . This corresponds to the apparent change in failure mode previously noted. Even with this change, however, the majority of the axially-aligned fibre tows in the impact Loaded specimen of fig. 9c still failed in tension. Thus the tensile strength of woven reinforced GFRP would not seem to be solely determined by the rate-dependent f r a c t u r e strength of the glass. In the present c a s e the extra factor which may determine the response at the highest r a t e s of strain i s the e a s e with which fracture3 fibre tows may debond from the matrix over considerable distances from the point of failure s o that frac- tures a c r o s s individual fibre tows a t widely separated sites in the specimen may nevertheless contribute to the final overall failure of the specimen. .The other significant difference between the present and the e a r l i e r results for woven reinforced GFRP is in the effect of strain r a t e on the strain to fracture. Results for the r a t e dependence of the fracture strain a r e compared in fig. 14 for the various types of composite. Within the experimental scatter ~f for the satin- weave CFRP specimens i s independent of strain r a t e over the entire range, to 103/s, while for satin-weave KFRP E~ shows a slight decrease from ~ 2 . 5 % to 1.2%. Both types of GFRP, however, show an increasing ~f with strain rate. For,fine plain-weave material the increase is modest, from%]. 8% to ~ 2 . 7 % but f o r the c o a r s e satin-weave specimens it i s dramatic, from
~ 3 % in quasi-static tests to%6.5% a t intermediate r a t e s while at impact r a t e s s+9% and in the one test f o r which it could be determined was found to be %13%. In both types of GFRP the extent of damage to either side of the f r a c t u r e surface increased with increasing strain r a t e but the effect was more marked for the coarse satin-weave reinforcement geometry. Since, in principle, a fine plain-weave fabric should be capable of a g r e a t e r extension on straightening than one of coarse satin-weave construction, the higher fracture strains in the Latter must be due, at Least in part, to the g r e a t e r extent of the damage zone. However, even if the axially-aligned fibres in the satin-weave fabric were to fully straighten under tensile load, the maximum possible ex- tension resulting would still only give a strain of %lo%. In the impact tests, therefore, the over- all failure strain must include a shear contribution resulting from pull-out of fibre tows. This implies that the near constant load in impact tests on satin-weave GFRP at strains greater than 2% i s a measure of the force required to pull out the fibre tows at these very high rates.
In the light of the foregoing discussion the following general description of the fracture behaviour of the satin-weave GFRP specimens i s proposed. At low and intermediate rates, where c l e a r tensile fracture surfaces a r e observed, the tensile strength is determined by the rate-dependent fracture strength of the glass fibre tows and most fibre fractures contributing to final failure a r e in a restricted region of the specimen close to the resulting fracture surface. At impact rates, however, where no c l e a r tensile fracture surfaces a r e observed, the fracture behaviour i s con- trolled by pull-out of fibre tows leading to a high value of fracture strain. The tensile strength
- Carbon
-.-. Kevlar
---- Glass
/I
Plain wave (ref. 11)
Fig. 10 Fig. 1 1
STRAIN-RATE DEPENDENCE OF TENSILE MODULUS STRAIN-RATE DEPENDENCE OF TENSILE MODULUS (present investigation) (woven GFRP specimens)
0
600 13- Fracture straln 1%)
O 0
_("A
1 2 - (1 result 1- B
: 400- ---'- --.-.-.-.-., /
E 11 - Present results
m
300 - -x- GFRP
ti 10- -0- KFRP
- 'g w
200 - - Carbon -0- CFRP
C Fevlar 9 - Ref 11
I00 - -A- Plan-weave GFRP
8 -
0 I I I I I I I
loL lo3 1cr2 lo-) lo0 lo1 lo2 103
Stra~n rate- log Fig. 12
STRAIN-RATE DEPENDENCE OF TENSILE STRENGTH (CFRP and KFRP specimens
- 1000- L -
2 E
j 800- 5 m
1 -
200 - Satln weave
0 I I I I I I I I 0 - I I I I I I I
loL lo3 IIJ~ la1 18 lo1 lo2 10) I'O-' 16) KI' 10' 10' lo2 10)
Stra~n rate - l o g l s ~ ' l Stra~n rate - I log (s-']
0 l I 1 1 1 1 1 1
laL 10\u2 101 lo0 lo1 lo2 1b3
I I I I I I I
Fig. 13 Fig. 14
STRAIN-RATE DEPENDENCE OF TENSILE STRENGTH STRAIN-RATE DEPENDENCE OF FRACTURE STRAIN (woven GFRP specimens)
lo-2 10.1 lo0 10' lo2 lb3 Strain rate -log 15-l) Strain rate- log (<')
C5-414 JOURNAL DE PHYSIQUE
may still be determined by the rate-dependent f r a c t u r e strength of the glass fibre tows but, be- cause such fractures occurring almost anywhere in the specimen may contribute to the final failure, on average the weaker tows will all fail f i r s t giving a lower overall fracture strength than at inter- mediate r a t e s where such fractures a r e Limited to a more restricted region of the specimen. To account for the change in failure mode at impact r a t e s it i s assumed that debonding between frac- tured fibre tows and the matrix becomes e a s i e r with increasing s t r a i n rate. This is not unreason- able since polymeric r e s i n matrices show a tendency to greater brittleness with increasing strain rate.
VI - CONCLUSIONS
Tensile stress-strain curves have been obtained at strain r a t e s of s 1 0 - ~ / s , s10/s and sl000/s for polyester resin specimens reinforced with the s a m e 5-end satin-weave fabric using carbon, Kevlar o r glass fibres. A closely similar rate-dependence of the tensile modulus i s obtained for all three composites, supporting a previous suggestion that the rate-dependence of the tensile modulus in woven-reinforced composites derives f r o m the elastic interaction between the rein- forcement and the resin matrix and i s determined by the rate-dependence of the matrix strength.
A significant increase in tensile strength at impact r a t e s in CFRP (and KFRP) tests, however, does not support e a r l i e r suggestions that the tensile strength of CFRP i s determined solely by the rate-independent f r a c t u r e strength of the carbon fibres and r a i s e s the possibility that, f o r the woven reinforcement configuration, the matrix may c a r r y a not insignificant proportion of the total load. F o r the GFRP specimens a change in the fracture mode i s observed, from that con- trolled by the rate-dependent fracture strength of the glass at Low and intermediate r a t e s to that controlled by resistance to pull-out of fibre tows a t impact rates, the change in mode resulting from a g r e a t e r tendency f o r debonding b e b e e n fractured fibre tows and the resin matrix a t the highest r a t e s of strain.
ACKNOWLEDGMENTS
Grateful acknowledgment is made to Fothergill and Harvey Ltd. f o r the supply of the composite materials, and to the Science and Engineering Research Council, for the award to one author (L. M. Welsh) of a r e s e a r c h studentship.
REFERENCES
/1/ Adams, D. F. and Miller, A. K., J. Mater. SCi. 11 (1976) 1697
/2/ Marom, G., Fischer, S., Tuler, F. R. and Wagner, H. D., ibid. 13 (1978) 141 9
/3/ Rotem, A. and Lifshitz, J. M., Proc. 26th Annual Tech. Conf. SPI Reinforced Plastics/
Composites Division (Society of Plastics Industry, New York, 1971) paper 10G /4/ Hayes, S. V. and Adams, D. F., J. Testing and Evaluation, 10 No.2 (1982) 61
/5/ Kawata, K., Hondo, A., Hashimoto, S., Takeda, N. and Chung, H.L., Proc. Japan-US Conf.
on Composite Materials (Japan Society for Composite Materials, Tokyo, 1981) 2 /6/ Ewins, P . D . , R.A.E. Technical Report No. 71217 (1971)
/7/ Bai, Y. and Harding, J., Proc. Int. Conf. on Structural Impact and Crashworthiness, (Imperial College, London, 1984) 2 482
/8/ Harding, J., Conf. on Mech. Props. at High Rates of Strain (Institute of Physics, London, 1979) 318
/9/ P a r r y , T. and Harding, J., Colloque International du CNRS No. 139 (OUEL Report No.
1365, 1981)
/lo/ Harding, J. and Welsh, L. M., Proc. 4th Int. Conf. on Composite Materials (Japan Society for Composite Materials, Tokyo, 1982) 845
/ l l / Harding, J. and Welsh, L. M., J. Mater. Sci. 18 (1983) 181 0
