ARTICLE IN PRESS
Composite Structures xxx (2006) xxx–xxx
www.elsevier.com/locate/compstruct
Low velocity impact induced damage evaluation and its effect on
the residual flexural properties of pultruded GRP composites
Z.Y. Zhang *, M.O.W. Richardson
Advanced Polymer and Composites (APC) Research Group, Department of Mechanical and Design Engineering,
University of Portsmouth, Portsmouth, Hampshire PO1 3DJ, UK
Abstract
Low velocity impact induced non-penetration damage in pultruded glass fibre reinforced polyester (GRP) composite materials was
investigated using an instrumented falling weight impact test machine with a chisel shaped impactor. The characteristics of the impact
event, force/time and force/deflection traces were determined. The internal damage was visualised and quantified by Electronic Speckle
Pattern Interferometry (ESPI) in terms of the thickness, density and uniformity degradations of fringe patterns. There is a linear relationship between the impact energy and the identified damage areas. The post impact structural integrity of impacted specimens was
evaluated by three point bending tests. It reveals that there is a significant reduction in flexural properties due to the impact-induced
damage and that the residual flexural strength is more susceptible to damage than residual modulus.
2006 Elsevier Ltd. All rights reserved.
Keywords: Polymer matrix composites (PMCs); Defects; Non-destructive testing; Impact behaviour
1. Introduction
Fibre reinforced polymer composites exhibit distinct
properties such as high specific modulus and strength,
lightness, high productivity, environmental degradation
resistance and cost effectiveness [1,2]. They have found
ever-increasing applications as engineering components
and structures in the fields of transportation, aircraft, aerospace, marine, military weapon system and nuclear energy
industries. However, polymer composites can be susceptible to impact damage during manufacturing and in service.
Unlike traditional materials such as metals and ceramics,
polymer composites exhibit unique damage characteristics.
When they are subjected to impact loading, there might be
no damage indication on surfaces by visual evaluation but
internal damage may have already occurred. This damage
can have an adverse effect on material performances and
structural integrity. It is referred to as Barely Visible
Impact Damage (BVID) [3–5]. This could potentially lead
*
Corresponding author. Tel.: +44 23 92842360; fax: +44 23 92842351.
E-mail address: [email protected] (Z.Y. Zhang).
to catastrophic failure at any subsequent moment. Therefore, it is important to develop a good understanding of
impact-induced damage non-destructively. Electronic
Speckle Pattern Interferometry (ESPI) has made a significant contribution to engineering measurement and testing.
It is capable of making whole-field and non-contact measurements of static and dynamic displacements with high
sensitivity [6–9]. It can be extended to damage visualisation
and quantification in non-destructive way [10–12]. The
presence of either external or internal damage leads to
mechanical degradations of GRP composite materials,
which can be visualised by ESPI in terms of uniformity
and continuity degradation and disruption of fringes. It
is also important to establish a relationship between damage severity and residual properties by which the residual
load bearing capabilities can be estimated [13–16]. The
objectives of this study are to develop a good understanding of the characteristics of low velocity impact induced
damage, including the damage formation and development
features, to visualise and quantify the impact induced damage using ESPI and to evaluate the post impact load bearing capabilities.
0263-8223/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compstruct.2006.08.019
Please cite this article as: Z.Y. Zhang, M.O.W. Richardson, Low velocity impact induced damage evaluation and its effect on ..., Composite Structures (2006), doi:10.1016/j.compstruct.2006.08.019
ARTICLE IN PRESS
2
Z.Y. Zhang, M.O.W. Richardson / Composite Structures xxx (2006) xxx–xxx
2. Experimental
The pultruded GRP panels were provided by Euro-Projects (LTTC) Ltd. They have been used for bridge floors,
marine and chemical plant decking where they are subjected to low velocity and non-penetration impact events.
The lay-up was confirmed to be symmetrical by deplying
the specimen. It consisted of roving in the centre, sandwiched with random glass fibre mat with woven fibre fabric
and surface veil placed symmetrically on both sides. The
fibre volume fraction was approximately 50%. The panel
was cut into rectangular specimens with overall dimensions
of 150 · 40 · 4 mm so that they could be accommodated
into a purposely build fixture for ESPI testing.
Impact testing was implemented using an Instrumented
Falling Weight Impact Test (IFWIT) machine. The general
features of the equipment are schematically shown in
Fig. 1. A controlled impact is achieved by dropping a chisel
shaped striker which is attached to a weight in a defined
manner from prescribed impact heights as shown in
Fig. 2. The probe is equipped with a pizeo-electric transducer for measurement of velocity and load interaction
between the probe and specimen. The velocity and load
transducer provides a complete profile of the deformation
response of the specimen during the impact process. Discrete values provided by the load transducer are collected
and stored in a computer for post processing. A pneumatically operated mechanism switches on when the impactor
bounces back to prevent the repetitive strike on the specimen. The specimen was pneumatically clamped between a
65 mm circular ring and anvil. Parameters such as energy,
time duration, load, deflection and velocity are generated
to characterise the impact events.
An ESPI system was based on a Helium Neon (He–Ne)
continuous wave laser (k = 632.8 nm) interferometer in
combination with a frame store, monitor, video recorder,
clamping jig and excitation apparatus which are schematically illustrated in Fig. 3. Theoretical aspect of ESPI and
detailed descriptions of the testing system are documented
elsewhere [17]. ESPI is capable of making whole-field and
non-contact measurements of deformations with high sensitivity and perturbations to the fringe patterns are corre-
Striker length: 15 mm
Striker radius: 2.5 mm
Impact test configuration
Diameter of supporter: 10mm
Diameter of loading roller: 10 mm
Span: 80 mm
Flexure test configuration
Fig. 2. Impact and flexure test configurations.
Laser
Mirror
Mirror
Beam Splitter
Lens
Lens
Digital correlation
Camera
Impacted Sample
Fig. 3. Schematic of electronic speckle pattern interferometry (ESPI).
Fig. 1. Schematic of instrumented falling weight impact test machine.
Please cite this article as: Z.Y. Zhang, M.O.W. Richardson, Low velocity impact induced damage evaluation and its effect on ..., Composite Structures (2006), doi:10.1016/j.compstruct.2006.08.019
ARTICLE IN PRESS
Z.Y. Zhang, M.O.W. Richardson / Composite Structures xxx (2006) xxx–xxx
3. Results and discussion
3.1. Characteristics of low velocity impact events
The impact performance of GRP composite materials
depends upon many factors, including the nature of the
fibre, matrix, interface, geometry and testing conditions.
Impact testing is generally split into two categories. The
first is impact testing under relatively low incident energy
where the composites are partially damaged but still capable of performing their primary function to certain extent.
The second is impact testing under high incident energy
where composites are completely ruptured or penetrated
by the striker. The former is generally simulated using a
low velocity falling weight or a swing pendulum tester
and the latter using a gas gun or some other ballistic
launcher. The responses of composites to impact loading
are extremely complicated and have already received considerable attention [19–22].
An investigation into the fracture toughness and impact
resistance of GRP composite materials was not the main
concern of this study. Instead, damage evaluation and
the establishment of correlation between the various
impact event parameters with post impact load bearing
capabilities were the primary concern. In order to achieve
these objectives, an instrumented low velocity falling
weight impact machine was employed to introduce damage
using a chisel shaped striker and the impact energy levels
were pre-determined and controlled by varying the striker
heights in order to incrementally introduce non-penetration damage in six specimens.
Force/time traces at six different impact loading levels
are illustrated in Fig. 4. They are indicative of load history,
which is associated with the damage initiation and development. The force/time traces of six impact events exhibit
similar trends with a complex series of peaks and troughs.
The overall curves can be approximately divided into two
parts at the maximum impact force. The first part is predominantly related to damage initiation. There is a linear
increase of force with the time at the start of loading, indicating the purely elastic response of the specimen and no
damage is expected to occur if the impact loading is terminated at this stage. When the impact loading in further
increased, there are pronounced fluctuations in damage initiation regions. It suggests that the damage initiation
occurs at different moments in different mechanisms. The
second part is mainly associated with damage propagation
in which the force has a monotonical decrease with increasing the time. It should be noted that force/time traces are
similar to those in penetration impact testing but the implications are fundamentally different. In the case of penetration impact events, load drops are associated with major
fracture processes such as fracture of fibre bundles or complete plies. As far as non-penetration impact events are
concerned, the load drops are associated with both elastic
and plastic deformation. The striker bounces upwards
from the specimen surface due to elastic collision and is
then caught by a second strike prevention mechanism.
Force/deflection traces at six loading levels are illustrated in Fig. 5. They are indicative of the deformation
and damage development history of specimens, which in
turn have an analogous relationship with force/time traces.
In general, force/deflection traces can be approximately
7
6.4 J
8.9 J
11.5 J
14 J
16.6 J
19 J
6
5
Force (kN)
lated to the impact induced barely visible damage in this
study.
The specimens were firmly clamped onto a purpose
build jig in which the rectangular specimen was clamped
vertically at both ends for ESPI measurement. Care was
taken to make sure that specimens were clamped in parallel
to the jig for easy calibration and analysis. Matt paint was
spread uniformly on the surfaces in order to enhance the
reflectivity. The specimens were mechanically excited using
a force transducer hammer. Live fringe patterns were
recorded for post damage recognition and measurement
[18].
Flexural modulus and strength of the impacted specimens were evaluated using a three-point-bending test.
The testing was carried out according to the ISO 178, a test
method for flexural strength and modulus of fibre reinforced plastics. The selected crosshead speed was 2 mm/
min so that specimens failed within 30–180 s intervals as
required. The experimental results were normalised in
order to achieve a straightforward interpretation and comparison of the post impact flexural strength and modulus.
The orientation of the damage with respect to geometry
of the flexure test specimens is illustrated in Fig. 2. Here
the normalised flexural modulus is defined as the modulus
of the sample experiencing the impact loading divided by
the modulus of the sample without experiencing the impact
loading. Similarly, the normalised flexural strength is
defined as the strength of the sample experiencing the
impact loading divided by the strength of the sample without experiencing the impact loading.
3
4
3
2
1
0
0
0.5
1
1.5
2
2.5
3
3.5
4
Time (ms)
Fig. 4. Force/time traces of six impact events with incremental incident
energies.
Please cite this article as: Z.Y. Zhang, M.O.W. Richardson, Low velocity impact induced damage evaluation and its effect on ..., Composite Structures (2006), doi:10.1016/j.compstruct.2006.08.019
ARTICLE IN PRESS
4
Z.Y. Zhang, M.O.W. Richardson / Composite Structures xxx (2006) xxx–xxx
7
6.4 J
8.9 J
11.5 J
14.0 J
16.6 J
19.0 J
6
Force (kN)
5
4
3
2
1
0
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
Deflection (mm)
Fig. 5. Force/deflection traces of six impact events with incremental
incident energies.
subdivided into different stages. The first stage is referred to
as linear load/deflection deformation due to the elastic
response to loading. No damage occurrence was expected
at this stage. A complete recovery from the deformed state
to the original specimen position is expected when the
applied load is removed. With the increase of load, the
specimen yielded, implying the onset of plastic or permanent deformation, lead to a sudden drop in load.
Once the load reached its maximum, it decreased monotonically and the deflection decreased accordingly due to
the gradual recovery of elastic deformation. Force/deflection traces of non-penetration impact events are substantially different from those of penetration impact events in
which the decrease of force is related to further damage
development when the striker passes through the specimen.
The maximum depth corresponds to the termination of the
impact event. In the case of non-penetration impact event,
the peak forces approximately coincide with the maximum
deflections in the force/deflection curves which are significantly different from the penetration impact events. The
deflections are then contracted due to the elastic deflection
recovery.
The summarised impact testing results are shown in
Table 1. Peak forces, peak energy, energy at failure and
peak deflection increased linearly with increasing the striker height. The peak deflections exhibited similar trends
which are characterised by a monotonical increase with
the incident energy. However linearity substantially deteriorates in deflections at failure because damage develop-
ment and elastic recovery in these specimens has an effect
on this process in complicated ways that make it difficult
to produce quantitative descriptions.
Energy is the primary parameter for characterising
damage tolerability, which is preferably employed to correlate with damage magnitude. It is inappropriate to correlate incident energy with damage magnitude because
part of this energy is transformed into kinetic energy
when the striker bounces upwards and dissipated in the
form of vibration and noise during the testing process.
These energy components have no direct relevance to
damage initiation and development. Peak energy consists
of the energy absorbed via elastic deformation of the
specimen and the energy dissipated via damage initiation
and propagation. For non-penetration impact events,
peak energy is invariably higher than the energy at failure
due to elastic recovery when the striker bounces upwards.
It is different from penetration impact events when peak
energy is invariably lower than energy at failure due to
the complete rupture of specimens. It is obvious that failure energy is exclusively responsible for the damage initiation and propagation in the specimen. In this case it is
significant to establish a correlation between energy at
failure and damage magnitudes and to evaluate postimpact load bearing capabilities.
The force/time and force/deflection traces can assist in
understanding of impact damage initiation and propagation mechanisms. However, they hardly indicate the impact
damage presence and severity. Under these circumstances,
ESPI was employed in this study for impact induced damage evaluation and quantification.
3.2. Damage evaluation
The representative features of impacted coupons on
both front and back surfaces are shown in Fig. 6. There
is usually no indication of damage presence on front sur-
Table 1
Impact testing results at incremental incident energy levels
Incident energy (J)
Striker height (mm)
Peak force (KN)
Peak deflection (mm)
Peak energy (J)
Deflection at failure (mm)
Energy at failure (J)
6.4
250
2.69
2.1
4.75
1.1
4.01
8.9
350
3.15
2.7
6.80
1.2
5.57
11.5
450
3.77
3.3
8.79
1.8
7.25
14.0
550
4.39
3.6
10.95
2.0
8.84
16.6
650
4.60
3.9
12.98
2.0
10.79
19.0
750
5.33
4.2
15.05
2.2
12.38
Fig. 6. Visual inspection of impacted coupons.
Please cite this article as: Z.Y. Zhang, M.O.W. Richardson, Low velocity impact induced damage evaluation and its effect on ..., Composite Structures (2006), doi:10.1016/j.compstruct.2006.08.019
ARTICLE IN PRESS
Z.Y. Zhang, M.O.W. Richardson / Composite Structures xxx (2006) xxx–xxx
face, where the specimen is subjected to compressive loading. It implies no defects in the impacted specimen which is
incorrect and misleading. By contrast, a crack can be visible on back surface, indicating the presence of damage. But
this indication is invariably unhelpful due to the physical
constraints and difficulties in accessing to the back surface
such as tubular, cylindrical and hollow components. Visual
inspection provides a certain indication of the damage
presence but it can hardly provide complete and quantitative information because of its inability in detecting internal damage.
Correlation fringe patterns generated by ESPI in combination with a force transducer hammer were employed to
characterise the internal damage. The typical fringe patterns for the sample without damage are characterised by
good symmetry and uniformity, which are associated with
the rigid body movement of the sample under excitation
loading. An interpretation of the fringe patterns requires
5
some knowledge in laser interferometry and observations
of live fringe pattern during the test would make the damage identification and evaluation less demanding. The
fringe patterns associated with six impact energy levels
are illustrated in Fig. 7. These images have been processed
to eliminate background noise and to highlight the damage
features. The damage is characterised by the degradation
and disruption of uniformity and continuity of fringe orientation, density and thickness. Damaged areas deform
more than un-damaged ones because of stiffness differences. The deformation in damaged area is plastic in nature due to the debonding, fibre pullout, delamination and
fibre breakage. Therefore the fringe patterns are frozen
because the damage is unlikely to recover elastically. Comparatively the deformation in undamaged area is elastic so
the fringe patterns would gradually disappear because of
the elastic deformation. Although these differences can
be very small in magnitude, ESPI can pick them up due
Fig. 7. Damage visualisation by fringe patterns from ESPI.
Please cite this article as: Z.Y. Zhang, M.O.W. Richardson, Low velocity impact induced damage evaluation and its effect on ..., Composite Structures (2006), doi:10.1016/j.compstruct.2006.08.019
ARTICLE IN PRESS
Z.Y. Zhang, M.O.W. Richardson / Composite Structures xxx (2006) xxx–xxx
to its high sensitivity. It can be seen that the magnitude
and severity of damage increase with increasing incident
impact energies.
There are lineal relationships between the damaged
areas and impact energies as illustrated in Fig. 8. The horizontal error bands illustrate standard variations of damage areas of five testing samples evaluated by ESPI.
Linearity between the damaged areas and energy at failure
was expected because it represents the energy dissipated or
absorbed at the expense of damage occurrence. The similar
linearity between the peak energy and the damaged areas
implies that a similar percentage of energy is absorbed in
the form of elastic deformation, which subsequently recovers elastically. It is expected that the damage related to different impact loading levels could be approximately
predicted although an extrapolation might not be completely appropriate. A wider impact stressing range is recommended for future work in order to develop a further
understanding of correlations between impact energy and
damage magnitude.
3.3. Residual flexural strength
Investigations into post-impact load bearing capabilities
of GRP materials involving different modes of stressing
have received a lot of attention. It is of particular interest
to understand to what extent the impacted materials can
sustain further loading. It has been found that tensile, compression and flexural properties are reduced when impact
damage is present in specimens [23,24]. In this study flexural testing was carried out to evaluate post impact properties, namely flexural strength and modulus. It was
anticipated that a correlation between the impact energies,
the damage magnitudes and residual flexural properties
would be established.
The relationships of residual flexural properties with
increasing impact damage energies are illustrated in
Fig. 9. It indicates that impact events result in reductions
in modulus and strength to varying degrees. The residual
flexural modulus is slightly less reduced with increasing
1
Residual Modulus
0.95
Residual Strength
Residual Properties (%)
6
0.9
0.85
0.8
0.75
0.7
0.65
0
2
4
6
8
10
12
14
Damage Energy (J)
Fig. 9. Residual flexural properties as a function of absorbed impact
energy.
impact energies compared to residual strength. The maximum reduction is approximately 20%, indicating that flexural modulus is not particularly sensitive to the damage
presence. Comparatively flexural strength is more severely
affected by impact damage, leading to higher reductions.
The maximum flexural strength reduction is approximately
30%. This higher sensitivity can be explained by the fact
that the impact damage is localised in most cases and therefore it has less effect on global properties such as modulus.
The calculation of modulus only involves the initial linear
part of force/deflection curves. Conversely localised impact
damage has adverse effect on the load bearing capability of
materials, referred to as ‘‘residual strength’’ or ‘‘strength
after impact’’. The flexural strength is calculated using
the ultimate force that the specimens could sustain and this
is dramatically reduced due to the presence of impact degradation. Similar experimental results were reported
regarding post-impact performance [25,26]. The relationship between the residual properties and the damaged areas
is similar to that in Fig. 9 and should be considered in the
context of the linear relationship between impact energies
and the damage areas as shown in Fig. 8.
4. Conclusion
16
Energy at Failure
14
Peak Energy
Impact Energy (J)
12
10
8
6
4
2
0
0
200
400
600
800
1000
1200
1400
Damage Area (mm2)
Fig. 8. A relationship between impact energy and damage area.
A series of samples were incrementally impacted using
an instrumented low velocity impact machine to introduce
non-penetration damage and were then subjected to NDT
and flexural testing. Force/time and force/deflection traces
are indicative of the damage initiation and propagation.
The impact events are characterised by the variations of
force, energy and deflection. ESPI has been successfully
employed to visualise and quantify the impact-induced
damage. Post impact flexural property testing shows that
flexural strength is more sensitive to the presence of localised impact damage than flexural modulus. The damage
areas increase with an increase in the energy at failure.
The successful evaluation of impact-induced damage using
ESPI could be extended to investigation of different
mechanical loading induced damage and performance
Please cite this article as: Z.Y. Zhang, M.O.W. Richardson, Low velocity impact induced damage evaluation and its effect on ..., Composite Structures (2006), doi:10.1016/j.compstruct.2006.08.019
ARTICLE IN PRESS
Z.Y. Zhang, M.O.W. Richardson / Composite Structures xxx (2006) xxx–xxx
degradations. The established correlations between impact
energy levels, damage severities and residual flexural properties can be employed to estimate post-impact load bearing capabilities for material design, selection and
maintenance purposes.
References
[1] Shalin RE. Polymer matrix composite. London: Chapman & Hall;
1995.
[2] Murph J. The reinforced plastics handbook. UK: Elsevier Science
Publishers Ltd; 1995.
[3] Cantwell WJ, Morton J. The significance of damage and defects and
their detection in composite materials. J Strain Anal 1992;27(1):
29–42.
[4] Mitrevski T, Marshall IH, Thomson R, Jones R, Whittingham B. The
effect of impactor shape on the impact response of composite
laminates. Comp Struct 2005;67(2):139–48.
[5] Poon C, Benak T, Gould R. Assessment of impact damage in
toughened resin composites. Theor Appl Fract Mechan 1990;13(2):
81–97.
[6] Steinbichler H, Gehring G. TV-holography and holographic interferometry: industrial applications. Opt Lasers Eng 1996;24:111–27.
[7] Genovese K, Lamberti L, Pappalettere C. A comprehensive ESPI
based system for combined measurement of shape and deformation of
electronic components. Opt Lasers Eng 2004;42:543–62.
[8] Farrant DI, Petzing JN, Tyrer JR. Geometrically qualified ESPI
vibration analysis of an engine. Opt Lasers Eng 2004;41:659–71.
[9] Hack E, Brönnimann R. Electronic speckle pattern interferometry
deformation measurement on lightweight structures under thermal
load. Opt Lasers Eng 1999;31:213–22.
[10] Leendertz JA, Butters JN. An image-shearing speckle pattern
interferometer for measuring bending moments. J Phys E: Sci
Instrum 1973;6:1107–10.
[11] Zhang ZY, Richardson MOW, Wisheart M, Tyrer JR, Pezting J.
Application of phase stepping ESPI to non-destructive testing of
GRP composite materials. J Brit Inst Non-destruct Test 1998;40(3):
183–7.
7
[12] Zhang ZY, Richardson MOW, Tyrer JR, Pezting J. ESPI nondestructive testing of GRP composite materials containing impact
damage. Comp Part A 1998;29(7):721–9.
[13] Huang JY. Prediction of the residual strength of laminated composites subjected to impact loading. J Mater Proc Technol 1995;54:
205–10.
[14] Kinsey A, Saunders DEJ, Soutis C. Post-impact compressive behaviour of low temperature curing woven CFRP laminates. Composites
1995;26:661–7.
[15] Short GJ, Guild FJ, Pavier MJ. Post-impact compressive strength of
curved GFRP laminates. Comp Part A: Appl Sci Manuf 2002;33:
1487–95.
[16] Zhang X, Davies GAO, Hitchings D. Impact damage with compressive preload and post-impact compression of carbon composite
plates. Int J Impact Eng 1999;22:485–509.
[17] Jones R, Wykes C. Holographic and speckle interferometry. Cambridge: Cambridge University Press; 1989.
[18] Zhang ZY, Ph.D Thesis, Loughborough University, 1999.
[19] Richardson MOW, Wisheart MJ. Review of low-velocity impact
properties of composite materials. Comp Part A 1996;27:1123–31.
[20] Lawcock G, Ye L, Mai YW. Progressive damage and residual
strength of a carbon fibre reinforced metal laminate. J Comp Mater
1997;31(8):762–87.
[21] Jang BZ, Hwang LC, Hawkes LR, Zee RH. The response of fibrous
composites to impact loading. Polym Comp 1990;11(3):144–57.
[22] Elder DJ, Thomson RS, Nguyen MQ, Scott ML. Review of
delamination predictive methods for low speed impact of composite
laminates. Comp Struct 2004;66:677–83.
[23] Bibo GA, Hogg PJ. Role of reinforcement architecture on impact
damage mechanisms and post impact compression behaviour. J
Mater Sci 1996;31:1115–37.
[24] Guild FJ, Hogg PJ, Richard JC. A model for the reduction in
compression strength of continuous fibre composites after impact
damage. Composites 1993;24(4):333–9.
[25] Lawcock G, Ye L, Mai YW. Progressive damage and residual
strength of a carbon fibre reinforced metal laminate. J Comp Mater
1997;31(8):762–87.
[26] Mouritz PA, Gallagher J, Goodwin A. Flexural strength and
interlaminar shear strength of stitched GRP laminates following
repeated impacts. Comp Sci Technol 1997;57(5):509–22.
Please cite this article as: Z.Y. Zhang, M.O.W. Richardson, Low velocity impact induced damage evaluation and its effect on ..., Composite Structures (2006), doi:10.1016/j.compstruct.2006.08.019