See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/272035902

The Finite Element Simulation of Lamb Wave Propagation in a Cracked

Structure with Coupled Filed Elements

Article  in  Advanced Materials Research · February 2012

DOI: 10.4028/www.scientific.net/AMR.463-464.618

CITATION READS

1 156

3 authors:

Samaneh Tashi Ali Abedian

Sharif University of Technology Sharif University of Technology
1 PUBLICATION   1 CITATION    64 PUBLICATIONS   773 CITATIONS   

SEE PROFILE SEE PROFILE

Romik Khajehtourian
ETH Zurich
30 PUBLICATIONS   132 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Reconfigurable Structures View project

Nonlinear wave propagation in dispersive media View project

All content following this page was uploaded by Romik Khajehtourian on 27 June 2016.

The user has requested enhancement of the downloaded file.

 The Finite Element Simulation of Lamb Wave Propagation in a Cracked
Structure with Coupled Filed Elements
Samaneh Tashi1, a*, Ali Abedian2, b, Romik Khajehtourian3, c
1, 2, 3 Department of Aerospace Engineering, Sharif University of Technology, Tehran 14646, Iran
a
samaneh.tashi@gmail.com, babedian@sharif.ir, cr.khajehtourian@gmail.com

Keywords: non-destructive evaluation, piezoelectric wafer active sensor, lamb wave, fem, crack
detection.

Abstract. Crack detection by Piezoelectric Wafer Active Sensors (PWAS) is one of the emerging
methods of Non-destructive Evaluation (NDE). These sensors can assess the health state of the
structure in far filed through the analyzing the high frequency Lamb wave propagation. As PWAS is
the essential part of this method, simulation and modeling of these sensors and their interaction with
the host structure, strongly affect the accuracy of results. In this study, unlike the previous works, in
which some certain areas of the host structure were considered as a sensor and actuator, the direct
simulation of electro-mechanical interaction of the PWAS and the host structure is modeled among
modeling the PWAS with a FEM coupled-field element. After verification of the FEM analysis with
theoretical group velocity of S0 mode, capability of the S0 mode in crack detection was examined.
Different length and angles of the crack in an aluminum plate modeled. The crack growth effect was
defined among a damage index by using the strain energy magnitude that received by sensor localized
behind the crack. The result showed that transmitted energy from the cracks decreases as the crack
length increases. In addition, as the angle crack to the sensor increases, the ability of sensor to crack
detection will increase.

Introduction
Many studies have contributed to crack damage detection based on Lamb wave propagation by
PWAS. Ihn and Chang [1] studied the hidden fatigue crack growth by using a network of piezoelectric
sensors, experimentally. In addition, Victor Giurgiutiu et al. have studied crack detection in metallic
and composite materials theoretically, experimentally and numerically and demonstrated the
capability of PWAS to detect damages in thin-wall structures [2,3,4]. Giurgiutiu et al. also studied
crack detection by Lamb wave propagation numerically through using coupled-field elements and
verified the results with the experiments, though their wok is restricted to a limited geometry and an
aluminum beam [5]. Ye Lu et al. quantitatively assessed the through the thickness crack size using
Lamb wave propagation in aluminum plates both numerically and experimentally. However, the
piezoelectric effect was not simulated directly and as a result, the interaction between mechanical
motions and electrical potentials were eliminated in a natural way [6]. Besides, in many studies, both
numerically and experimentally, the average crack length was larger than 8 mm. Therefore, the ability
of this method to detect cracks smaller than 8 mm has not been reported. Also, in many crack detection
with PWAS, the crack is generally considered perpendicular to the wave front and as a result the
interaction between Lamb waves and the crack is usually simplified into normal transmission and
reflection, only. However, with limited number of transducers and unknown crack position, the
incident Lamb waves are usually oblique to the crack and so the interaction of angled cracks with
PWAS sensors has been rarely investigated.
The quantitative relationship between cracks and Lamb wave scattering is usually defined through
a concept called Damage Index (DI). This concept which is related to the wave propagation strategy
can be described by the reflection or transmission energy in time or frequency domains [7]. This index
could be a function of the reflection or transmission energy divided to the pristine signal energy which
is propagated by actuator in time domain [8].
 Table 1. Geometry and mechanical properties of the host structure
Length Wide Thickness Elastic Modulus Poisson Ratio Density
(mm) (mm) (mm) (GPa) (Kg/m3)
500 500 1 73 0.3 2700
In this paper, the interaction of normal and oblique Lamb wave incident at 300 kHz frequency,
with a through the thickness crack of different length and angles in an aluminum plate by means of
Finite element Method (FEM) is evaluated. While in this study, the wave propagation strategy has
been pitch-catch method, DI is established as a function of transmitted energy divided by pristine
signal energy. Also, the influence of crack angle on ability of the sensors to detect the damage is
discussed.

Lamb Wave Interaction with a Crack

Ultrasonic NDE method can usually detect cracks with a characteristic size larger than one-half of the
wavelength of ultrasonic wave. The ultrasonic wave which can be guided through the thickness of the
plate and therefore detect the through the thickness cracks are Lamb waves. Lamb waves can exist in
symmetric and antisymmetric types. For each type there are several modes that their propagation
velocity depends on the frequency and the plate thickness product.
Since various Lamb modes simultaneously might excite in a structure at a frequency, making the
PWAS tune into one specific Lamb mode is necessary to prevent from complexity of wave behavior.
Therefore, the tuning frequency is a frequency at which all mode shapes decay to zero and just one
mode remains strong. Tuning frequency of each mode which is excited by a PWAS depends on
thickness of the bonding layer, geometry of the PWAS, thickness and material of the host structure
[9]. Even though the wavelength of A0 - the fundamental antisymmetric mode shapeis shorter than the
S0-the fundamental symmetric mode shape- the preferable Lamb wave mode for crack detection is S0
because this is much less dispersive than A0. Therefore, it can provide a strong reflection from through
the thickness cracks at low frequencies to interpret the received signals [2]. In accordance with the
ideal bonding layer, the geometry of the PWAS, thickness and material of the host structure which
will be described in Table 1 and Table 2, the tuning frequency for the exciting S0 mode is 300 kHz
(Fig. 1), [10].
Table 2. Geometry and mechanical properties of PWAS
Properties of PWAS
(kg/m3) 𝜌 7700
d33 ( 10-12 m/V ) 400
-12 -175
d31 ( 10 m/V )
d15 ( 10-12 m/V ) 590
Poisson ratio, v 0.35
63
Young Moduluse Y11EGPa
54
Young Moduluse Y33E GPa
Thickness(mm) 0.07
Sensor Size (mm2) 7×7
Modeling of Crack Detection Based on Lamb Wave Propagation with PWAS
A three dimensional FEM model was developed by a commercial FEM software to investigate crack
detection with PWAS. The PWAS were modeled directly with the coupled field elements and excited
by electrical signals using 3D eight-node brick solid piezoelectric element. Then the electrical
response of PWAS to the wave motion examined to assess the health state of the host structure. The
host structure was a 500 mm 500 mm aluminum plate in which the 3D eight-node brick solid
elements were employed with free-free boundary condition.
 Fig. 1. Predicted Lamb wave strain response on a 1-mm aluminum plate under a 7-mm PWAS
excitation [10].

The mechanical properties of the plate and electromechanical properties of the PWAS are listed in
Table1 and Table 2. Based on the wave propagation speed and in order to minimize the reflection
from the boundaries, actuator was localized at plate center and sensors were respectively localized at
x= 130 mm and y= 150 mm from the plate center. The element length was set to be 1 mm in plane
and two layers were modeled in the thickness direction to maintain a length–thickness ratio of 1.25,
which is sufficiently accurate to portray the characteristics of Lamb waves. A 300 kHz five count
smoothed tone burst (Fig. 2) was used to excite the structure with S0 Lamb wave mode. The electrical
voltage directly fed onto the top surface of the PWAS, simulating the operation as in a real application.
Fig. 3 shows the Lamb wave propagation in a pristine plate.

Fig. 2. 300 kHz five count smoothed tone burst Fig. 3. S0 Lamb wave propagation in a
pristine plate

In order to validate the model, the S0 has excited in a pristine plate and group velocity of simulated
wave compared to the theoretical S0 group velocity to examine the accuracy of the model. Then the
through-thickness cracks which are perpendicular to the wave front were simulated with length 4, 6,
8, 10 mm to examine the ability of this FEM method in crack detection. Also the influence of oblique
cracks on signals received by the sensors was investigated by simulating the 30, 60 and 90 degree
angled crack to the sensor.

Numerical Results
Results validation. Each numerical result needs to be verified with an experimental or theoretical
result to assure the reliability of the FEM analysis. The group velocity of a 300 kHz S 0 mode in an
aluminum plate with 1mm thickness is evaluated 5.44 Km/s by theory [10]. The received signal in
sensors in distance x=130 mm and y=150 mm from the actuator in a pristine plate has shown in Fig.
4.TOF of the first wave packet received in sensors x=130 mm and y=150 mm are respectively 40.6
and 44.6 micro sec and as a results the group velocity of each wave packet is 5.41 and 5.35 Km/s.
Fig. 5 shows the plot of distance Vs. time predicted by theory, and two FEM analysis points
corresponding to the sensors. The correlation between theory and FEM is very good, which gives
confidence in our investigation.

Fig. 4. Received signal in sensor x= 130 mm and y= 150 mm in a pristine plate

Crack Length Effect on Scattering Lamb Waves. The crack presence in a structure causes
discontinuity and the discontinuities strongly affect the wave propagation behavior in an elastic
medium. Compared with the amplitude of received wave at the same sensor in an intact plate and
cracked plate, the amplitude of transmitted Lamb wave decreases in the defective plate as the crack
length increases. The received signal by the sensor behind the crack, for crack length, 4, 6, 8 and 10
mm is shown in Fig 6a. A Damage Index to establish a quantitative relationship between crack length
and Lamb waves scattering is defined as follows:

𝒓𝒆𝒄𝒊𝒆𝒗𝒆𝒅 𝒆𝒏𝒆𝒓𝒈𝒚 𝒃𝒚 𝒕𝒉𝒆 𝒔𝒆𝒏𝒔𝒐𝒓 𝒊𝒏 𝒄𝒓𝒂𝒄𝒌𝒆𝒅 𝒑𝒍𝒂𝒕𝒆

DI 1 −
𝒓𝒆𝒄𝒊𝒆𝒗𝒆𝒅 𝒆𝒏𝒆𝒓𝒈𝒚 𝒃𝒚 𝒕𝒉𝒆 𝒔𝒆𝒏𝒔𝒐𝒓 𝒊𝒏 𝒑𝒓𝒊𝒔𝒕𝒊𝒏𝒆 𝒑𝒍𝒂𝒕𝒆

The DI will be zero if the structure has no defect and it goes to 1 as crack length increases. The
strain energy magnitude received by the senor, for different crack lengths is depicted in Fig. 7a. As a
result the DI for each crack length has been computed and represented in Fig. 8. The DI diagram
shows the expected trend: the index increases as the crack is getting larger.
Angled Crack Effect on Scattering Lamb Waves. In fact, the oblique incident scattering Lamb
wave is simulated through modeling the different angled cracks. Even though, all cracks have the
same length, as a crack angle to the sensor decreases, the ability of sensor to recognize the damage
decreases because in addition to reflection and transmission energy from the crack, the amount of
energy diffracted from the angled cracked and make the wave behavior complex and needs more
study. The amplitude and strain energy magnitude of the oblique cracks are shown in Fig. 6b and Fig.
7b respectively. The results show that to detect an angled crack, using the sensor array is necessary.

Fig. 5. Comparison of theoretical group velocity of S0 with FEM group velocity of S0

 Fig. 6. The received signal by the sensor behind the crack; (a) different crack lengths; (b)
different oblique cracks

Conclusion
In this paper, the possibility of using FEM to detect cracks in an aluminum plate with PWAS was
investigated. PWAS were directly modeled with coupled-field elements and excited by electrical
signals. The interaction of normal and oblique incident Lamb waves with through-thickness cracks in
different lengths through propagating the S0 mode was studied. The accuracy of model was verified
by theoretical results among comparison between FEM and theoretical S0 group velocity. A Damage
Index was established to define the quantitative relationship between crack length and Lamb wave
scattering. It is concluded that the transmission coefficient decreases with crack length increase. Also
it is demonstrated that to detect oblique cracks, using an array of sensors is necessary because in
addition to reflection and transmission, the incident Lamb wave will be diffracted.

Fig. 7. The strain energy magnitude received by the senor; (a) for different crack lengths; (b) for
different oblique cracks
 Fig. 8. DI for each crack length

References
[1] Ihn, J.B., Chang, F.K., 2002, “Multicrack Growth Monitoring at Riveted Lap Joints Using
Piezoelectric Patches,” in 7th Annual International Symposium on Non-destructive Evaluation
for Health Monitoring and Diagnostics, March 17–21, San Diego, CA, Proceedings of the SPIE,
Vol. 4702, 29–40
[2] Giurgiutiu, V., Bao, J., 2002, “Embedded Ultrasonic Structural Radar for the Non-destructive
Evaluation of Thin-wall Structures,” in Proceedings of the 2002 ASME International Mechanical
Engineering Congress, November 17–22, New Orleans, LA, paper IMECE200239017.
[3] Giurgiutiu, V., Cuc A., 2005, “Embedded Non-destructive Evaluation for Structural Health
Monitoring, Damage Detection, and Failure Prevention”, The Shock and Vibration Digest.
[4] Giurgiutiu, V., Yu, L., Kendall, J., Jenkins, C., 2007, “In situ imaging of crack growth with
piezoelectric wafer active sensors”, AIAA Journal, Vol. 45, No. 11, pp. 2758-2769.
[5] Liu, W., Giurgiutiu, V., 2007, “Finite Element Simulation of Piezoelectric Wafer Active Sensors
for Structural Health Monitoring with Coupled-Filed Elements” Proc. of SPIE Vol. 6529 65293R-
1,
[6] Lu, Y., Yu, L., 2008, “Quantitative assessment of through-thickness crack size based on Lamb
wave scattering in aluminum plates” Science Direct- NDT&E International 41 59–68.
[7] Diligent O, Grahn T, Bostrom A, Cawley P, Lowe MJS, 2002, “The low-frequency reflection and
scattering of the S0 Lamb mode from a circular through-thickness hole in a plate: finite element,
analytical and experimental studies”. J AcoustSoc Am; 112(6):2589–601.
[8] Demma A, Cawley P, Lowe MJS, Roosenbrand AG, (2003) “The reflection of the fundamental
torsional mode from cracks and notches in pipes”. J AcoustSoc Am;114(2):611–25.
[9] Santoni, G. B.; Yu, L.; Xu, B.; Giurgiutiu, V., 2007, “Lamb Wave Mode Tuning of piezoelectric
Wafer Active Sensors for Structural Health”, ASME Journal of Vibration and Acoustics, Vol.
129, No. 6, pp. 752-762.
[10] Giurgiutiu, V., 2008, Structural Health Monitoring with Piezoelectric Wafer Active Sensors,
Elsevier Academic Press, 760 pages, ISBN 978-0120887606.

View publication stats