Impact damage evaluation in woven composites using acoustic and thermoelastic techniques

Carlo Santulli 
 20
 
In other potential areas of application, the higher costs of the raw materials for the 
manufacture of high-tech composites (for example carbon, glass or Kevlar fibre 
reinforced polymers) seem at first not justified. Nevertheless, in these fields, the use of 
composites can offer substantial advantages. For example, consider the building 
industry. In some countries, where wood is extensively used in buildings, renewable 
forest resources are not sufficient to cover demand. This concern created interest in the 
development of materials for replacing timber. Ideally, these materials should be fire 
resistant, have excellent flame characteristics and be able to resist severe environmental 
conditions (high temperature gradients and high humidity) without suffering significant 
degradation.  
 
The need to reduce energy consumption as a result of environmental and economical 
concerns has increased the use of composite materials in the advanced transportation 
sector. Another consequence of these concerns is that there is a greater interest in 
developing recyclable composites, or even in obtaining them from waste materials, for 
example from used PET containers or jute sacks (1,2). This results from the fact that 
composites are considered to be a special waste by many countries and therefore 
complex and expensive operations are required for their recycling. 
In this context, there is a growing interest in the use of natural fibre reinforced 
composites. Natural fibres include straw, flax, hemp and jute (Fig.2), etc., and are 
already harvested in many countries of the world. Their primary advantage over wood is 
their renewal time of one year in contrast to approximately 30 years for softwood.  
Impact damage evaluation on woven composites using acoustic and thermoelastic techniques 
 21
 
 
Fig.2 Typical patterns of jute fabric used for carpets 
 
In the United Kingdom, AEA Technology in collaboration with the International Jute 
Organisation has undertaken a number of projects involving the use of jute composites. 
The aim of these projects was to find marketable applications for jute fibre composites 
in South East Asia and to replace traditional shrinking markets. In the mid-seventies, 
attempts to use jute composites to replace wood in village houses in a number of Third 
World countries have been reported (3). Hemp has also found use as a building material. 
In particular MDF (medium density fibre) hemp composite boards have a fibre density 
and stiffness comparable to that offered by wood. In addition, these materials can be, 
after suitable fibre treatments, completely impermeable to water. Non-toxic composite 
boards, based on hemp straw or coir fibres, have also been manufactured (4-6). Another 
possibility of using natural fibres is the production of totally biodegradable composites. 
Carlo Santulli 
 22
This can be achieved when the natural fibre (e.g., jute) is employed as a reinforcement 
of a biopolymer matrix. Combustion of biopolymer (e.g., Biopol®, a bacterial 
polyester) maintains carbon dioxide neutrality and is completely slag-free (7). 
 
Many studies have been performed to investigate the mechanical and structural 
characteristics of high-tech composites. As a result, current standards and guidelines are 
generally oriented towards carbon and glass fibre composites. Clearly, low-tech 
composites are also subjected to severe in-service conditions, and failure to fully 
characterise their behaviour might result in a reduction in product performance, 
rendering them less marketable. For example, a natural fibre composite floor covering 
should be able to resist home heating, withstand dropped objects and withstand people 
walking on it, etc., during many years of service.  
 
1.2 Impact testing 
 
Impact is an event that is frequently observed during the operation of a composite 
structure. Consider for example the front cab of a high-speed train, this can suffer 
damage from boulders or objects present on the rail, birds or other animals, hailstorms, 
etc. Even when the impact event creates non-visible damage in the material, its effect on 
the mechanical properties of the composite has to be taken into account. As a 
consequence, an evaluation of the possible effect of impact loading on the properties of 
a composite structure should be performed.  
Impact damage evaluation on woven composites using acoustic and thermoelastic techniques 
 23
 
Ideally, the impacted component should be examined to yield answers to four basic 
questions. These can be summarised as follows: 1. Is there any damage? 2. Where is it 
located? 3. What is its size? 4. Which failure processes are involved? In simple terms, 
we need to detect, localise, size and characterise the impact damage. Characterisation 
refers to the different types of damage induced by an impact in a composite material: 
most frequently this includes matrix cracking, delamination and fibre fracture.  
An important aspect of the damage assessment is the evaluation of damage criticality, 
that is, to assess whether an impacted structure can continue in service or not. To 
achieve this goal, damage detection and measurement are not sufficient and a simulation 
of the operating conditions is required. Knowledge of damage development during 
operational service should lead to a prediction of residual life, which will allow the 
engineer to guarantee that the structure will survive for a given time under normal 
operating conditions. This is often difficult to achieve, so in practice the design of the 
component tends to be conservative, in the sense that the safety factors are high enough 
to account for both the possible inaccuracy of the prediction of residual life and for 
unpredictable events occurring during service (8). Of course, this procedure involves 
higher costs. The costs of testing a composite structure can be compensated through the 
possible reduction of safety factors without adversely affecting the reliability of the 
structure. This reduction in safety factors would finally result in a reduction in 
production costs for the manufacturer. 
 
To determine the suitability of a particular composite for a given application, a series of 
tests on specimens is usually required before passing to the tests on a prototype 
Carlo Santulli 
 24
component. Here again, impact testing is widely used, often followed by a detailed 
inspection of damage. Post-impact mechanical tests are then required to measure 
material degradation due to impact. During these tests, a number of real-time 
monitoring techniques are available to support and integrate the information obtained 
when measuring the degradation of the mechanical properties. These techniques should 
not produce any further damage within the material, but only detect the effect of the 
applied stress; for this reason, these techniques are defined as passive monitoring 
methods. For structures, it is a frequent practice to combine some of these techniques 
together (9, 10), since a single technique may not be able to detect all of the damage 
present in the material. In addition, most techniques give relative rather than absolute 
indications of damage, so that a comparison between results obtained using different 
techniques on the same set of specimens can provide useful feedback on data reliability. 
These considerations explain why these techniques will never be able to fully replace 
direct observation of damage. 
Here, we focus our attention on two types of passive technique: acoustic emission and 
thermoelastic stress measurement. The former is based on the detection of ultrasound 
coming from stress wave propagation and the latter on temperature gradients induced by 
stresses in materials as a result of the thermoelastic effect. The information acquired 
regarding the effect of impact loading on the stress distribution during a mechanical test 
is then correlated with the real damage observed in the material. The final aim of this 
programme is to formulate a global method for the observation of damage in impact-
loaded composites by using and comparing the results from different techniques.  
Impact damage evaluation on woven composites using acoustic and thermoelastic techniques 
 25
CHAPTER 2 
LITERATURE REVIEW 
 
2.1 Introduction 
 
The impact of foreign objects on composite materials has been widely investigated in 
recent years (1-54). A large number of works concern impact modelling (3-13), damage 
monitoring (16, 17, 19-31) and the assessment of post-impact residual properties (32-
54). Clearly, it would be impossible to cover all of the published literature on impact, as 
it would largely exceed the aim of this work. Some introductory books are nevertheless 
given in references (1, 3-7). 
This review is opened by a general discussion on the impact characterisation (Section 
2.2) and the assessment of the post-impact properties of composites (Section 2.3). Some 
aspects of the mechanics of woven composites are then briefly introduced (Section 2.4) 
(55-71). Following this, the materials studied in this work are discussed (72-84). Glass 
fibre composites (Section 2.5) do not need more than a brief introduction, and their 
impact behaviour is well documented (1, 5-7, 71). In contrast, jute fibre reinforced 
composites (Section 2.6), although they are not strictly speaking a new material, have 
received less coverage in the literature, and are therefore more thoroughly presented and 
discussed in this thesis (85-92). A limited number of works also exist regarding their 
mechanical properties (93-96). 
A number of techniques for monitoring impact damage in composites are available in 
the literature (97-116) and details of these procedures can be found in Section 2.7. The 
work presented in this thesis concentrates on real-time techniques applied during post-
Carlo Santulli 
 26
impact tests. In particular, acoustic emission (Section 2.8) (117-178) and thermal 
techniques (Section 2.9) (179-246) are discussed in detail. Among thermal techniques, 
emphasis is given to the presentation of thermoelastic stress measurement (225-246).  
 
2.2 Impact characterisation 
 
2.2.1 General observations  
The impact process may involve relatively high contact forces acting over a small area 
for a period of short duration. Local strains generated at the point of contact between the 
two solids result in the absorption of energy. If the absorbed energy exceeds a threshold, 
the impact event may result in damage. When a projectile strikes a laminated composite, 
fracture processes such as delamination, matrix cracking and fibre fracture frequently 
occur. 
The study of the force acting locally as a consequence of the impact event is the aim of 
impact dynamics. The development of impact dynamics started with the establishment 
of a contact law (1), initially developed for the simple case of the impact of a beam by a 
steel sphere (2). The contact law provides a relationship between the contact force and 
the indentation, defined as the difference between the displacement of the projectile and 
that of the back face of the laminate. Starting from this approach with the aim of 
developing the analysis to study impact loading on composites, the effect of composite 
anisotropy cannot be neglected: in most cases however, the hypothesis of material 
orthotropy can be retained for a laminated composite (3, 4). 
The initial effect of impact loading is to produce some damage close to the surface of 
the laminate. In most cases however, the internal plies of the composite are damaged as 
Impact damage evaluation on woven composites using acoustic and thermoelastic techniques 
 27
well, and, as the projectile penetrates into the laminate, the material stiffness changes 
locally (5). Partly as a result of this, the contact force cannot be assumed to be constant 
for all of the duration of the impact event (6). From an observation of the indentation 
process on composite laminates, it was concluded that the loading and unloading phases 
exhibit different characteristics, and hence should be described by different laws (7). 
Moreover, in an attempt to reproduce real loading situations, different projectile 
geometries were employed in order to measure the modifications of the response of the 
composite (8). The results of such tests on plates are, however, not always applicable to 
structural design: some emphasis has therefore been placed on the need to account for 
scaling effects from plate to structure (9, 10). Further complications introduced during 
the testing of woven composites are briefly discussed in Section 2.6. The difficulty in 
overcoming these problems is reflected by the fact that the relationship between the 
contact force and the resulting indentation is generally determined experimentally on 
laminates. Semi-empirical relations are, nevertheless, employed in literature (11, 12). 
Simple models enable the designer to account for the effects of impact on a structure: an 
example of this is a simple mass-spring model for objects dropped on a composite 
laminate. This represents a common impact problem (13). In this quasi-static model, the 
impact response is represented by a time dependent force and the target composite is 
represented by an equivalent mass with equivalent stiffness (14). However, to obtain a 
good agreement between a model and the impact loading is sometimes difficult (4). A 
common problem is that impact loading can result in the generation of different forms 
of damage in the composite (5). All these forms of damage need to be considered, 
because all of them are likely to influence the residual mechanical properties of the 
material. For example the British Standards define five failure processes (these being: 
Carlo Santulli 
 28
crack, break, penetration, shattering and indentation). The standard therefore 
recommends two methods for carrying out impact testing on composites (staircase and 
statistical) both of which aim a specific level of damage to be incurred by the specimen 
during an impact (15). 
 
2.2.2 Low and high velocity impact loading 
The impact process is usually characterised by the kinetic energy or the velocity of the 
projectile. A low velocity impact does not always result in the perforation of the 
composite (16). Low velocity impact testing is often able to give indications of the 
extent and shape of damage due to real impacts occurring during service (17). In the 
case of high velocity impact, the perforation energy is in some cases referred to as the 
ballistic limit (18). Apart from the incident energy, the ballistic limit is a function of 
many other factors, including the target density, the impact angle and the target area (1). 
In the aircraft industry, in order to assess the continued airworthiness of the part 
subjected to impact loading, a BVID (Barely Visible Impact Damage) concept is 
frequently used. This criterion originated from periodical inspections, where it was used 
to give an initial indication of the area to be examined. A correlation was also found 
between the value of the BVID energy and the stiffness of the laminate (19). Between 
BVID and the perforation energy, a third threshold corresponding to the energy for the 
onset of rear face damage is often defined (20). 
In low velocity impact dynamics, semi-experimental models based on simple 
assumptions are quite common. For example, in some situations such as in sandwich 
beams (21), the response of the laminate to low velocity impact loading can be 
described using a two dimensional analysis, where the applied impact load is assumed 
Impact damage evaluation on woven composites using acoustic and thermoelastic techniques 
 29
to be uniform across the specimen width. Another common assumption is to exclude the 
effect of the degradation of material properties due to the impact event (4). However, 
when the contact force acting on the impacted material locally exceeds its elastic limit, 
such assumptions can no longer be maintained. The introduction of multiple impacts on 
the same specimen has also been carried out with the aim to select the best method (i.e., 
insertion of blade or T-stiffener) to improve the impact resistance of a composite panel 
(22). 
 
2.2.3 Impact tests and damage measurement 
Most impact tests are conducted using a falling-weight tower, in which the mass of the 
impactor is large and its velocity low, or using an air gun system with a small mass (for 
example a half-inch steel ball), which is propelled at a high velocity. The falling-weight 
test simulates the impact created by a tool being dropped on a structure, while the air 
gun tests simulate flying debris during the take-off and landing phases on aircraft 
operation (23). Typical contact durations are 60 µs for air gun impacts and 1500 µs for 
drop-weight impacts (19). To measure the energy required to penetrate the composite 
plate, a test, referred to as dart penetration test, is also used. In this test, a weighted, 
hemispherical head dart is dropped from an adjustable release mechanism at a 
predetermined height. A photoelectric speed trap then measures the free-falling dart 
speed and calculates the loss of kinetic energy, thereby determining the energy required 
to penetrating the laminate (15). 
 
A common method used to investigate the impact response of composites is to drop 
masses with different energies on the same set of specimens. This practice allows one to 
Carlo Santulli 
 30
evaluate the overall damage tolerance of the material and to measure the progressive 
degradation of its mechanical properties with increasing impact energy. To obtain 
impacts with different energies, either the height from which the mass is released, or the 
mass of the projectile is varied. Variables that are usually recorded during the test are 
the impact force, the acceleration of the projectile, its incident and rebound velocities 
and the deflection-time profile of the laminate. 
In the case of static indentation, the laminate is clamped in order to prevent buckling, 
whilst the indentor, which can be flat or hemispherical, impinges on the laminate. Static 
indentation is generally performed up to the damage threshold of the laminate or by 
increments up to a given indentor displacement. After the test, any permanent 
indentation is measured (24). 
 
For the case of drop-weight impact testing, models can be based on force, to model 
damage initiation (3) or on energy, to predict damage (25). This second approach is 
more suitable in studies dealing with the effects of a real impact, which may generate 
extensive damage resulting in a severe reduction of the tensile and compressive strength 
of the composite. Therefore many workers have studied the variation of damage area 
with increasing impact energy (26, 27). The irregular characteristics of this area make it 
necessary to approximate the damage to a flaw of well-defined size and shape for 
subsequent modelling of the reduction in mechanical properties of composites (26). 
One of the variables that are often measured during impact tests on composites is the 
laminate deflection. This parameter is measured because its value reflects the local 
stiffness of the laminate and clearly depends on material properties and laminate size. 
Depending on the lay-up, the deflection can range from one to several times the 
Impact damage evaluation on woven composites using acoustic and thermoelastic techniques 
 31
laminate thickness (3). Moreover, the extent of damage is dependent on the stiffness of 
the target; hence under low velocity impact conditions laminate bending will induce 
high tensile stresses in the lowest ply (27). Experimental studies to determine the 
contact force and strain histories at different locations in an impacted plate have shown 
that, for a given energy, the effects of transverse shear deformation are less important 
for drop-weight impact loading than for high velocity impact (16). Studies have been 
performed to correlate the deflection of the target with the size and mass of the impactor 
(28, 29) and to observe the effect of other conditions, such as absorbed moisture (30). 
Many studies have been conducted on impact-damaged laminates in order to establish 
prevalent damage patterns and shapes (31, 32). Delaminations, which can result in a 
reduction in ply stiffness, are generally only found between plies with different 
orientations. An effect of delaminations that was also investigated is the local buckling 
of the laminate (33). The shape of the delaminations changes with the orientation of 
plies and is generally circular (34) or elliptical (35). Individual delaminations may 
exhibit a butterfly wing-shaped (1) or twin-lobed peanut shaped shape (33, 36). In other 
cases, a distinction between circular, diamond-shaped, elongated and waisted 
delaminations was also done (37). Shear cracks (perpendicular to the impacted surface) 
and tensile cracks (inclined) are also frequently observed (5). Observing the nature of 
impact damage through the section of the specimen, it appears to be approximately 
conical (35). Damage has been therefore often referred to as pine tree or reversed pine 
tree (1) (Fig.3). 
Carlo Santulli 
 32
 
 
Fig.3 Schematic cross-section of an impacted laminate 
with tensile matrix cracks 
 
2.3 Post-impact mechanical tests  
 
Undertaking mechanical tests on composites allows one to evaluate their post-impact 
properties, which in turn are linked to the damage tolerance of the system. A range of 
different tests are frequently conducted in order to achieve this aim: the most commonly 
used tests are residual tensile strength after impact (RTSAI), compression after impact 
(CAI), shear, buckling and fatigue tests and less commonly residual flexural strength 
testing. 
Impact damage evaluation on woven composites using acoustic and thermoelastic techniques 
 33
Post-impact tensile testing has been performed in a number of studies to investigate the 
effect of impact-induced damage on the degradation of the mechanical properties in the 
composite (38-40). With this aim, an impact energy threshold has been often measured, 
beyond which the tensile properties of the composite are likely to fall below that 
required by the designer (5). Here, the ultimate stress and strain of the fibres are primary 
factors in determining the residual tensile properties of the composite (41). In contrast, 
the influence of the matrix has been shown to be less important in a number of works 
(27, 42). 
The results of residual tensile strength after impact (RTSAI) tests can be also employed 
in models relating the damage area to the stress concentration induced by a circular 
hole. Comprehensive models capable of accounting for all of the failure processes (fibre 
failure, matrix cracking and delamination) have also been proposed, but their correlation 
with experimental data has proved more difficult (43). 
 
Generally, the compression-after-impact (CAI) (Fig.4) test procedure is based on a 
rectangular plate, in which the vertical sides are supported, and a compression load is 
applied to the horizontal edges. The severity of this method for damage evaluation in 
impacted composites is highlighted by the fact that strength reductions measured by the 
CAI test are greater than in the other mechanical tests. Compressive strength reduction 
is due to the delamination areas produced in the laminate as an effect of impact and was 
shown to depend on both stacking sequence and surface treatment of the laminate (44). 
Indeed, it has been suggested that the basic tools for a composite engineer dealing with 
impact problems should be a drop-weight tower, an ultrasonic C-Scan and a CAI test rig 
(45). Impact damage produced by a mass striking a composite can be initially observed