18th World Conference on Nondestructive Testing, 16-20 April 2012, Durban, South Africa

Evaluation of Fatigue Damage in Composites with Various Defects Using

Air-coupled Guided Waves

Martin RHEINFURTH 1, Frank SCHMIDT

2, Richard PROTZ

3,

Gerhard BUSSE 1, Peter HORST

2, Maik GUDE

3, Werner HUFENBACH

3

1 Institute for Polymer Technology (IKT), University of Stuttgart, Stuttgart, Germany;

Phone: +49 711 685 62669, Fax: +49 711685 62066;

email: martin.rheinfurth@ikt.uni-stuttgart.de, gerhard.busse@ikt.uni-stuttgart.de 2 Institute of Aircraft Design and Lightweight Structures (IFL), Technical University of Braunschweig,

Braunschweig, Germany; frank.schmidt@tu-bs.de, p.horst@tu-braunschweig.de 3 Institute of Lightweight Engineering and Polymer Technology (ILK), Technical University of Dresden,

Dresden, Germany; r.protz@ilk.mw.tu-dresden.de, m.gude@ilk.mw.tu-dresden.de, ilk@ilk.mw.tu-dresden.de

Abstract

Based on mode conversion from ultrasound in air to guided waves, tubes made of glass fibre reinforced plastic

are investigated in a single-sided access configuration. Fatigue damage is induced by biaxial mechanical loads in

specimens with and without defects. Using the proposed non-contact methodology, porosity, low velocity impact

damage, and fibre waviness are detectable before cyclic loading. During fatigue loading, damage formation

resulting in ultimate failure and effects of defects are monitored by air-coupled guided wave area scans. The

observed decrease in guided wave velocity correlates closely to stiffness degradation of the composite due to

cyclic loading. Increase in guided wave attenuation corresponds to development of crack density being visible in

the transparent composite. Early initiation of delamination due to porosity and fibre waviness are detected using

the non-contact ultrasound setup. The initiation spot of ultimate failure is in many cases predictable depending

on existence and type of pre-damage.

Keywords: Guided waves, composite, fatigue, effects of defects, fibre waviness, impact damage, porosity, air-

coupled ultrasound

1. Introduction

In times of sustained globalisation and growing environmental problems, resource-conserving

mobility and renewable energies gain in importance. Rising demand for transportation of

passengers and goods raises growth of civil aerospace and similar industries. However,

arising scarcity of fossil energy sources and essential climate protection force reduction in

structural weight to minimize fuel consumption. This reduction can be accomplished using

fibre reinforced polymer (FRP) because of its high specific stiffness and strength. The

advantages of FRP including simplified manufacturing processes also explain its application

in the area of renewable energies, as production of rotor blades for wind energy plants.

Increasing usage of FRP challenges non-destructive testing (NDT). Detection of the large

variety of defects occurring in heterogeneous and anisotropic polymer-matrix composites

during their manufacturing process and operation requires advanced non-destructive

evaluation (NDE). For example, fatigue damage of metals is characterised by nucleation and

growth of a single crack, which is detectable using conventional penetrant or eddy current

testing. In comparison, fatigue mechanisms of (frequently non-conducting) FRP involve many

different defects and more complex accumulation of damage. NDT of composites is highly

relevant because fatigue damage induced by typical cyclic loading of aircrafts or wind

turbines reduces residual strength and stiffness of structural components. Detection and

monitoring (during service) of manufacturing or operation induced defects is important to

evaluate their detrimental effect on structural integrity and fatigue resistance.

After an overview on fatigue damage and defect kinds in composites, a short literature survey

on exemplary NDT methods for damage evaluation is presented. This paper focusses on the

principle of air-coupled guided waves which is outlined referring to existing literature.

Excerpts of recent research outcomes obtained by using this technique are presented.

2. Fatigue damage and types of defects in composites

Mechanical induced defects in composites can be categorised in delamination, matrix crack,

and fibre fracture. A delamination in FRP is characterised by debonding of two plies with

different fibre orientations. Matrix cracks often occur as transverse cracks: a single ply is

prone to crack through its entire thickness parallel to the fibres due to mechanical loading

orthogonal to its reinforcement direction. Failure of a large number of filaments in one ply is

considered as fibre fracture.

Reifsnider [1] distinguishes between three stages of fatigue damage evolution in FRP under

tension fatigue loads which also applies for tension-compression loads [2]. The first stage is

characterised by formation of transverse cracks resulting in a steep drop in Young’s modulus

of the FRP. The second stage of fatigue damage evolution involves crack consolidation and

development of small delaminations on the tips of transverse cracks which is accompanied by

only little stiffness degradation. Due to formation of large delaminations and fibre fracture,

stiffness rapidly decreases in the last stage of fatigue life till ultimate failure.

Bird strike, hailstorm, and tool drop are classic hazards for structural components made of

FRP. The resulting impact damage is often undetectable by visual inspection of the outer

component surface even though an impact may result in large internal delaminations and

decrease in fatigue life [3]. An unavoidable feature of large FRP parts is porosity which has a

detrimental effect on fatigue performance depending on void distribution and void volume

content [4]. Besides porosity, also excessive fibre waviness is a common manufacturing

induced defect leading to early formation of delamination and ultimate failure under fatigue

loading [5].

3. Review on NDT methods for various defects in FRP

The geometry of a delamination in composites can be readily characterised by various NDT

techniques such as thermography with optical excitation [6], speckle interferometry [7], defect

resonance [8], conventional ultrasound c-scans, and many more. Single matrix cracks and

very small delaminations, however, are hardly detectable using the above mentioned

methodologies. Thermography with ultrasonic excitation detects matrix cracks close to the

specimen surface [9]. Penetrant enhanced x-ray radiography can be used for detection of

matrix cracks and small delaminations [10]. However, radio-opaque liquid penetrates only in

cracks connected to the surface, which limits inspection to laminates with only a few layers

and consolidated cracks. Reduction in ultrasonic amplitude measured in pulse-echo

configuration correlates with increase in matrix crack density [11]. The attenuation of

ultrasound waves is also used to determine the volume content of voids [12]. Fibre fracture

and delamination events can be monitored using acoustic emission [13]. Characterisation of

fibre waviness in FRP components is a challenging issue, which is tackled using ultrasonic

waves [14] and terahertz NDE [15].

Lamb waves velocities in plate-type FRP components can be used to monitor stiffness

degradation induced by fatigue loading [16]. Attenuation measurements of circumferential

guided waves in tubes excited by contact transducers correlate with crack density and stiffness

degradation [17]. Using air-coupled ultrasonic transducers, Lamb wave assessment can be

performed in a non-contact and single-sided configuration [18]. Area scans with air-coupled

Lamb waves are suited to detect delaminations [19]. This paper summarizes and compares

recent investigation in the area of air-coupled guided waves for fatigue monitoring involving

FRP tubes with and without pre-damage [20-22]. Furthermore, preliminary research results

involving porosity are presented [23]. The differences in area scans of typical defects as

impact damage, fibre waviness, and void accumulation are discussed. The effect of

mechanical fatigue loading on the results of area scans is introduced. The correlation of

guided wave velocity and attenuation, Young’s modulus degradation, and matrix crack

density is illustrated by an exemplary specimen without pre-damage. The effect of defects on

this correlation is outlined referring to the literature [20-23].

4. Experimental approach

Filament winding and resin transfer moulding (RTM) was used to manufacture tubular

specimens made of glass fibre reinforced plastic (GFRP) [20]. Fibre waviness was introduced

by low thread tension during the winding process, which led to deviation of position and

orientation of the rovings (fibre bundles) close to the outer surface of the tube [21]. Variation

of the RTM process parameters and air injection are used to accomplish evenly distributed

and accumulated porosity, respectively [23]. For the investigation of impact related pre-

damage a polyamide projectile was shot on the specimen [22]. Fatigue damage was

introduced in the GFRP tubes (wall thickness: 2 mm; diameter: 46 mm) by a servo-hydraulic

biaxial testing machine. This allows for the study of damage formation resulting from a

combination of reversed normal loads (tension-compression) and reversed shear loads

(negative-positive torsion). Young’s modulus of the tubes was calculated from data obtained

by the hydraulic testing machine [20].

The specimens were repeatedly removed from the test rig after a certain number of cycles for

NDE. Commercial air-coupled ultrasound transducers with a centre frequency of 207 kHz and

an active diameter of 8 mm were used to excite guided waves in the tubes [20]. The generated

guided mode is akin to the lowest antisymmetric plate wave (a0-mode). The ultrasound

radiated from the guided wave was detected by a receiving transducer to measure phase and

amplitude. In order to calculate phase velocity and attenuation, the propagation distance of the

guided wave was varied by moving the receiver away from the stationary transmitter. Area

scans of the tube wall were performed by keeping the distance of the two transducers constant

while rotating and translating the tubular specimen. Crack densities, delamination shapes, and

porosity distribution were characterised by visual inspection of the transparent GFRP [20-22].

4. Comparison of result excerpts for tubes with and without defects

Thick rovings as compared to wall thickness and resin accumulations between the rovings

cause a quite uniform stripe pattern in the area scans for nominal flawless regions of the tube

[20]. This consistent pattern of straight stripes is disturbed in areas of fibre waviness (Figure

1). Over the full length of the tube between 90° and 240° of the circumference, patterns of

wavy and interrupted stripes indicate deviation of roving positions and orientations. During

fatigue loading, this results in early and excessive delamination along wavy rovings which

was detected by area scans and confirmed by visual inspection [21]. High speed imaging

revealed that always one of the observed delaminated areas leads to ultimate failure.

Size and shape of impact-induced delamination is definable by air-coupled area scans because

of high amplitude deviation and missing stripe pattern (Figure 2). Even the depth distribution

can be estimated by phase area scans [22], which was asserted by visual inspection. Fatigue

loading resulted in some cases in growth of delamination being detectable by area scans

before all specimens started to fail in the area of the impact damage.

Figure 1. Area scan of tube with fibre waviness before

fatigue loading (colour bar indicates deviation in per

cent from arithmetically averaged amplitude) [21]

Figure 2. Area scan of tube with impact induced

delamination before fatigue loading [22]

Areas of high void concentration can be identified by sudden intensity and width variations of

the typical stripes (Figure 3, position: 60 mm, 120°). In contrast to the observed wavy stripes

caused by fibre waviness, however, the stripes remain straight. Latest results imply that

critical void accumulations, which lead to delamination during cyclic loading, appear in the

area scan. The onset of catastrophic failure was always located in one of the delaminated

areas. Very recent results indicate that specimens with finely distributed porosity display

features which look very much like those obtained on nominal defect-free specimens. These

porosity specimens also exhibit similar fatigue behaviour. More detailed results will be

presented in an upcoming journal article [23].

Fatigue damage of nominal flawless tubes includes development of spots with accelerated

degradation process which are detectable using thermography during cyclic loading [20] and

air-coupled guided waves between the fatigue steps (Figure 4). The deviation image between

the area scans after 95.6% and 98.7% reveals three of those spots. Comparison of area scans

disregards the inherent inhomogeneity of the GFRP (stripe pattern) and shows more clearly

the evolution of fatigue damage than a single air-coupled scan. A high speed camera disclosed

that the most obvious spot (position: 120 mm, 100°) initiated final failure of the specimen

[20].

The guided wave attenuation rises by about 40% throughout fatigue life in the presented data

(Figure 5). However, more data published in [20] shows that the heterogeneity of this

particular filament winding composite and resulting inhomogeneous fatigue evolution affects

the attenuation results. The increase in attenuation is caused by scattering on matrix cracks

and small delaminations. Within the first 15% of loading cycles till final failure, 90% of

matrix cracks develop (Figure 5). In addition to crack consolidation, the formation of

delamination on tips and intersections of transverse cracks explains the incline in attenuation

even after crack saturation. The rapid formation of matrix cracks corresponds to the steep

drop in Young’s modulus and guided wave velocity in the first 15% of fatigue life (Figure 6).

Gradual decline in velocity and stiffness till ultimate failure of the tube are in good

agreement. Since locally limited accelerated fatigue evolution insignificantly affects the

overall stiffness of the comparable large specimen, progressive stiffness degradation just

before final failure (as reported in [1, 2]) is absent. More experimental data about damage

evolution monitored by air-coupled guided waves including various loading conditions is

presented in [20].

Figure 3. Area scan of tube with void accumulation

before fatigue loading

Figure 4. Amplitude deviation in per cent of the area

scans at 95.6% and 98.7% of fatigue life (nominal

flawless tube) [20]

Compared with nominal defect-free GFRP tubes, less fatigue damage was observed for

specimens with fibre waviness or impact damage till final failure. Outside of the pre-damaged

areas, matrix crack formation is interrupted before saturation by early ultimate failure of the

specimen. The same applies to stiffness degradation and decrease in guided wave velocity

[21, 22]. The effect of the investigated porosity is currently being studied. Porosity with low

volume content seems to have only a small influence on the investigated fatigue indicators

[23].

Figure 5. Attenuation and normalised crack density

over fatigue cycles of the nominal flawless tube (data

from Rheinfurth et al. [20])

Figure 6. Guided wave velocity and Young’s modulus

over fatigue cycles of the nominal flawless tube (data

from Rheinfurth et al. [20])

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.2 0.4 0.6 0.8 1.0

No

rma

lise

d v

alu

es

Load cycles over cycles till ultimate failure

Guided wave attenuation (normalised byinitial value)

Crack density (normalised by crack densityat saturation after 40% of fatigue life)

0.86

0.88

0.90

0.92

0.94

0.96

0.98

1.00

1.02

0.0 0.2 0.4 0.6 0.8 1.0

No

rma

lise

d v

alu

es

Load cycles over cycles till ultimate failure

Phase velocity of guided wave(normalised by initial value)

Young's modulus (normalisedby initial value)

5. Conclusions

NDE based on air-coupled guided waves allows for detection and characterisation of

accumulated voids, (impact-induced) delamination, and excessive fibre waviness in GFRP

tubes. These three different types of pre-damage are readily distinguishable in single-sided

access configuration by means of area scanning. Guided wave area scans can also monitor

evolution of fatigue damage unaffected and affected by initial defects. If not hidden by

formation of a large delamination, the initiation spot of ultimate failure is most often

predictable towards the end of the specimen’s lifetime. The air-coupled configuration with

variable transducer distance is well suited for velocity and attenuation measurements of

guided waves even in rather heterogeneous composites. The observed phase velocity and

attenuation data is a useful measure of fatigue damage including stiffness degradation and

matrix crack density. The presented non-contact methodology, which is most likely applicable

to all kinds of FRP and shell-type structural components, is a versatile indicator for

manufacturing and operation induced damage.

Acknowledgements

The authors gratefully acknowledge the support by the German Research Foundation (DFG)

as part of the project PAK267.

References

1. K L Reifsnider and K Schulte, 'Long-term fatigue behavior of composite materials', In T

K O'Brien, Long-term Behavior of Composites, pp 136-139, ASTM STP 823, 1983.

2. K Schulte, 'Compressive static and fatigue loading of continuous fibre-reinforced

composites', In S E Groves and A L Highsmith, Compression response of composites

structures, pp 278-305, ASTM STP 1185, 1994.

3. K W Kang and J K Kim, 'Fatigue life prediction of impacted carbon/epoxy laminates

under constant amplitude loading', Composites Part A, Vol 35, No 5, pp 539-535, 2004.

4. A R Chambers, J S Earl, C A Squires, and M A Suhot, 'The effect of voids on the

flexural fatigue performance of unidirectional carbon fibre composites developed for

wind turbine applications', International Journal of Fatigue, Vol 28, No 10, pp 1389-

1398, 2006.

5. G B Murri, 'Influence of ply waviness on fatigue life of tapered composite flexbeam

laminates', In P Grant and C Q Rousseau, Composite Structures: Theory and Pratice, pp

188-209, ASTM STP 1383, 2000.

6. A Gleiter, C Spießberger, and G Busse, 'Lockin-thermography with optical or

ultrasound excitation', 10th

International Conference of the Slovenian Society for Non-

Destructive Testing Proceedings, September 2009.

7. P Menner, P Schmitz, H Gerhard, and G Busse, 'Lockin-speckle-interferometry with

modulated optical and inductive excitation', 10th

ECNDT Proceedings, June 2010.

8. I Solodov, J Bai, S Bekgulyan, and G Busse, 'A local defect resonance to enhance

acoustic wave-defect interaction in ultrasonic nondestructive evaluation', Applied

Physics Letters, Vol 99, No 21, 2011.

9. W Hufenbach, M Gude, R Protz, G Busse, Ch Spiessberger, M Rheinfurth, 'Influence of

impact related predamage on the material behavior of composites under high dynamic

tensile loading' INTECH Proceedings, September 2010.

10. W W Stinchcomb, 'Nondestructive evaluation of damage accumulation processes in

composite laminates', Composites Science and Technology, Vol 25, No 2, pp 103-118

1986.

11. S Tsushima and M Ono, 'Nondestructive evaluation of fatigue damages in FRP using

ultrasonic waves', US-Oacific Rim Workshop on Composite Materials for Ship and

Offshore Structures Proceedings, April 1998.

12. I M Daniel, S C Wooh, and I Komsky, 'Quantitative porosity characterization of

composite materials by means of ultrasonic attenuation measurements' Journal of

Nondestructive Evaluation, Vol 11, No 1, pp 1-8, 1992.

13. R S Williams and K L Reifsnider 'Investigation of acoustic emission during fatigue

loading of composite specimens' Journal of Composite Materials, Vol 8, pp 340-355,

October 1974.

14. K Y Kim, W Zou, and W Sachse, 'Wave propagation in a wavy fiber–epoxy composite

material: Theory and experiment', Journal of Acoustic Society of America, Vol 103, No

5, pp 2296-2301, 1998.

15. R F Anastasi, 'Investigation of fiber waviness in a thick glass composite beam using

THz NDE', SPIE Proceedings, Vol 6934, 2008.

16. M D Seale, B T Smith, and W H Prosser, 'Lamb wave assessment of fatigue and thermal

damage in composites', Journal of the Acoustical Society of America, Vol 103, No 5, pp

2416-2424, 1998.

17. S Adden, K Pfleiderer, I Solodov, P Horst, G Busse, 'Characterization of stiffness

degradation caused by fatigue damage in textile composites using circumferential plate

acoustic waves, Composites Science and Technology', Vol 68, No 7-8, pp 1616-1623,

June 2008.

18. M Castaings and P Cawley, 'The generation, propagation, and detection of Lamb waves

in plates using air‐coupled ultrasonic transducers', Journal of the Acoustical Society of

America, Vol 100, No 5, pp 3070-3077, 1996.

19. I Solodov, D Döring, 'Ultrasonics for NDE of fiber-composite materials', In: Damage

and its evolution in fiber-composite materials: Simulation and non-destructive

evaluation, Ed.: G Busse, B-H Kröplin, F K Wittel, pp 17–36, ISD-Verlag, 2006.

20. M Rheinfurth, F Schmidt, D Döring, I Solodov, G Busse, and P Horst, 'Air-coupled

guided waves combined with thermography for monitoring fatigue in biaxially loaded

composite tubes', Composites Science and Technology, Vol 71, No 5, pp 600-608,

March 2011.

21. F Schmidt, M Rheinfurth, P Horst, and G Busse, 'Effects of local fibre waviness on

damage mechanisms and fatigue behaviour of biaxially loaded tube specimens',

Composites Science and Technology, Available online 19 September 2011.

22. F Schmidt, M Rheinfurth, R Protz, P Horst, G Busse, M Gude, and W Hufenbach,

'Monitoring of multiaxial fatigue damage evolution in impacted composite tubes using

non-destructive evaluation', Composites Part A: Applied Science and Manufacturing,

Available online 8 December 2011.

23. F Schmidt, M Rheinfurth, P Horst, and G Busse, 'Multiaxial fatigue behaviour of GFRP

with evenly distributed or accumulated voids monitored by various NDT

methodologies', Submitted to International Journal of Fatigue.