Title: Structural Health Monitoring with Piezoelectric Wafer Active Sensors
Author: Victor Giurgiutiu
Paper presented at:
16th International Conference of Adaptive Structures and Technologies ICAST-2005,
10-12 October 2005, Paris, France
ABSTRACT
Piezoelectric wafer active sensors (PWAS) are inexpensive, non-intrusive un-obtrusive
devices that can be surface-mounted on existing structures, or inserted in a new composite
structure. The PWAS can be used in both active and passive modes. In active mode, the PWAS
generate Lamb waves that can exist as either traveling waves or standing waves. As traveling
waves, PWAS-generated Lamb waves can be used with the pitch-catch, pulse-echo, or phasedarray methods that arrow far-field and some medium-field damage detection. As standing
waves, PWAS-generated Lamb waves can used in conjunction with the electro-mechanical
(E/M) impedance technique that allows near field and some medium-field damage detection.
This paper presents new results obtained in the use of PWAS for the structural health
monitoring of aerospace vehicles. One set of results to be presented will refer to experiments
performed on detection the crack propagation in a fracture-mechanics panel with a stressconcentration precrack in the center. The crack was propagated using cyclic fatigue loading in a
large testing machine. The crack was imaged with the PWAS EUSR method, which is the
guided Lamb wave equivalent of the P-wave C-scan, only that it is produced from a single
location and does not need a raster x-y coverage like the conventional P-wave C- scan. PWAS
EUSR measurements were taken while the testing machine was running, thus simulating inservice monitoring conditions. Another set of results refers to the PWAS monitoring of
disbonds in an spacecraft panel. The E/M impedance technique is shown to distinguish clearly
between a bonded and disbonded part. The E/M impedance of bonded part measured at different
locations is shown to be consistently the same, with the actual curves overlapping, thus
confirming that the impedance changes noticed at the disbonded location are solely due to
disbonding.
PWAS
Piezoelectric wafer active sensors (PWAS) are inexpensive, small, and unobtrusive
transducers that operate on the piezoelectric principle by coupling the electrical and mechanical
energy fields (Figure 1). For embedded NDE applications, PWAS couple their in-plane motion
with the Lamb-waves particle motion on the material surface. The in-plane PWAS motion is
excited by the applied oscillatory voltage through the d31 piezoelectric coupling. PWAS act as
both exciters and detectors of ultrasonic Lamb-wave through in-plane strain coupling.
Embedded ultrasonic NDE is achieved by using the pitch-catch, pulse-echo, and phased-array
NDE methods with in-situ PWAS transducers. A comprehensive study of the state of the art in
embedded ultrasonic NDE was published by Giurgiutiu and Cuc (2005).
PWAS
~
V(t)
t = 2h
PWAS array
2 in (50 mm)
S0
λ /2
(a)
Figure 1
(b)
(a) PWAS array on an aircraft panel; (b) typical structure of S0 Lamb waves in
interaction with PWAS
10
9
8
Volts (mV)
7
6
5
4
3
2
1
A0
S0
0
0
50 100 150 200 250 300 350 400 450 500 550 600 650 700
Freq (KHz)
(a)
Figure 2:
(b)
Lamb-wave tuning using a 7-mm square PWAS placed on 1.07-mm 2024-T3
aluminum alloy plate: (a) experimental results; (b) prediction with Equation (1) for
6.4 mm effective PWAS length.
1.1 SELECTIVE TUNING OF LAMB WAVE MODES WITH IN-SITU PWAS
We have developed closed form solutions to predict how the Lamb waves are excited by a
surface mounted PWAS transducer (Giurgiutiu, 2003). These solutions have been developed
using space-domain Fourier transform of the Lamb-waves differential equations and of the
shear excitation given by the PWAS transducer. The Fourier transform equations were shown to
accept closed-form solutions in the case of “ideal bonding” between the PWAS and the
structure. Since actual bonding may differ from the ideal case, the ideal-bonding closed form
solutions can be taken as an upper bound on the results. For a rectangular PWAS and a straightcrested Lamb-wave front, the closed form solution takes the form:
S
A
aτ 0
aτ 0
S N S (ξ ) i (ξ x −ω t )
A N A (ξ ) i (ξ x −ωt )
ε x ( x, t ) = −i
sin ξ a
e
i
sin
ξ
a
e
−
(1)
∑
∑
µ ξ
µ ξ
DS′ (ξ S )
DA′ (ξ A )
where τ (ξ ) is the space-domain Fourier transform of the shear stress coming from the PWAS,
i.e.,
τ ( x) = aτ 0 [δ ( x − a) − δ ( x + a ) ] , τ (ξ ) = aτ 0 [ −2i sin ξ a ]
(2)
In Equations (1) and (2), ε x ( x, t ) is the direct strain in the structure at the interface with the
PWAS transducer, a is the half-length of the PWAS, τ 0 is the amplitude of the shear stress at
S
S
A
A
the interface, ξ is the wave number, ω is the circular frequency, N (ξ ) and D(ξ ) are closed form
solution functions that appear during the solution process, while the subscripts S and A denote
the symmetric and antisymmetric Lamb wave modes. (For full details of this derivation, see
Giurgiutiu, 2005). Raghavan and Cesnik (2004) extended our work to circular PWAS and
circular crested Lamb waves, and obtain the following Bessel functions solution:
ε r (r , t ) z = d = π
τ 0 a iωt ⎡
N (ξ S ) ( 2) S
N (ξ A ) ( 2) A ⎤
e ⎢ ∑ J1 (ξ S a )ξ S S
H1 (ξ r ) + ∑ J1 (ξ A a)ξ A A
H1 ( ξ r ) ⎥
µ
ξ
⎢⎣ ξ
⎥⎦
DS ′ (ξ S )
DA′ (ξ A )
S
(3)
A
Equations (1) and (3) present opportunities for tuning into different Lamb-wave modes
depending on the PWAS geometry and excitation frequencies. We have developed MATLAB
programs to predict such tuning curves and then verified these predictions against carefully
conducted experiments. Figure 2 shows the results for a 7-mm square PWAS placed on 1.07mm 2024-T3 aluminum alloy plate. The experimental results (Figure 2a) indicates a rejection of
the highly dispersive A0 Lamb wave mode at around 200 kHz, when only the less dispersive S0
mode is excited. This “sweet-spot” is beneficial for pulse-echo damage detection. On the other
hand, a strong excitation of the A0 mode is observed at around 50 kHz. We found that these
experimental results are perfectly reproduced by Equation (1) if we take the effective PWAS
length as 6.4 mm (Figure 2b). The difference between the actual PWAS length and effective
PWAS length can be attributed to shear transfer/diffusion effects at the PWAS boundary.
Similar good agreement was obtained for circular PWAS and Equation (3). Further work done
on thicker plates (3.15 mm thick 7075 alloy) revealed again good agreement with the theoretical
model even in the presence of several Lamb-wave mode. Full details of these experiments and
MATLAB simulations are given by Bottai and Giurgiutiu (2005).
Guided Lambwave C-scan
Threshold Controls
Threshold
controls
Dial to change angle and
find different peaks for AScan angles
Large reflection from crack
Figure 3:
A-scan for a given
dial angle
The PWAS EUSR graphical user interface permit the adjustment of the A-scan
angle that generates the rays in the guided Lamb-waves C-scan
PWAS EUSR METHOD
PWAS can be configured in phased arrays. Hence, the phased array principles can be used to
image large structural areas from a single location. These ideas were applied by Giurgiutiu and
Bao (2004) to image cracks in large plates using an array of eight 7-mm square PWAS. They
named this concept embedded ultrasonics structural radar (EUSR). Subsequently, we have
performed extensive work on achieving in situ imaging of crack growth using various phased
array of PWAS transducers and the EUSR concept. We have developed specific Lamb-waves
phased array data analysis algorithms that have been applied to the design and analysis of
several test configurations including large plates, thick parts, and curved panels. Full details of
this work are given by Yu and Giurgiutiu (2004, 2005) and Giurgiutiu, Jenkins and Cuc (2005).
Figure 3 shows the PWAS EUSR graphical user interface. The upper right corner presents a
guided Lamb-wave C-scan performed through an azimuth sweep from the PWAS phased array
location. This C-scan is the guided-wave equivalent of a P-wave scan. However, its major
advantage is that it can be performed from a single location.
FRACTURE PROPAGATION – IN-SITU IMAGING OF CRACK GROWTH WITH
PWAS PHASED ARRAYS
To illustrate the capabilities of the PWAS EUSR method we perform a laboratory test on a
aluminum plate subjected to cyclic fatigue loading. With this experiment, we aimed at verifying
the capability of the PWAS EUSR method to perform in-situ detection of structural cracks
while the structure is subjected to cyclic loads, i.e., similar to an actual in-service measurement.
We also wanted to verify that the PWAS EUSR method is capable of monitoring the crack
growth while the structure is subjected to in-service cyclic loads.
Crack
180 mm
Crack
711 mm
PWAS
phased array
PWAS array
(a)
Figure 4:
(b)
597 mm
Experiment for crack growth imaging with PWAS phased arrays and the EUSR
algorithm: (a) test specimen in the fatigue machine; (b) schematic diagram.
Optical image
PWAS EUSR image
Crack length and cycles
30-mm at 0 kilocycles
(precrack)
35-mm after 22
kilocycles
47-mm after 40
kilocycles
50-mm after 42
kilocycles
55-mm after 48
kilocycles
60-mm after 58
kilocycles
Figure 5:
Crack growth imaging results with PWAS phased arrays and the EUSR algorithm
showing correlation between optical and PWAS EUSR images (data was taken
while testing machine was running cyclic loading to simulate in-flight recording)
Finally, we wanted to calibrate the imaging obtained with Lamb-wave PWAS EUSR method
against a direct optical imaging performed with a digital camera. Figure 4 shows the fatigue
crack-growth setup. A 1-mm thick 2024-T3 plate was mounted in an MTS 810 fatigue test
system and subjected to cyclic fatigue loading. The plate contained an initial precrack of 30 mm
length. The plate was cycled at appropriate loads to promote accelerated crack growth. The
specimen was instrumented with a PWAS phased array. Readings were taken on-line during the
actual fatigue testing using standard laboratory equipment and some purpose-built electronics.
Over 58 kilocycles fatigue test duration, the crack grew from the initial 30 mm to a final 60 mm
length. Seven crack growth increments were recorded (30, 40, 47, 50, 55, and 60 mm). Very
good agreement between PWAS phased array EUSR imaging and the optical images was
obtained (Figure 5).
E/M IMPEDANCE RESULTS
The electromechanical (E/M) impedance method is a damage detection technique
complementary to the wave propagation techniques. The mechanical impedance method
consists of exciting vibrations of bonded plates using the PWAS as a transducer that
simultaneously transmitter and receiver of elastic waves. The effect of a piezoelectric wafer
active sensor affixed to the structure is to apply a local strain parallel to the surface that creates
stationary elastic waves in the structure.
Tektronix TDS 210
oscilloscope
Panel 1
3.0
A
0.190 Fastener, Csk (FS)
A
0.125
(a)
5:1
Sect A-A
Figure 6
0.060
24.0
3.0 dia
HP 33120
Signal generator
23.5
Note: All dimensions in inches
(b)
Spacecraft test panel tests: (a) structural design (NextGen Aeronautics, Inc.); (b)
experimental setup
Through the mechanical coupling between the PWAS and the host structure, on one hand,
and through the electro-mechanical
80
transduction inside the PWAS, on the
a1
a2
a3
70
other hand, the drive-point structural
impedance is directly reflected into
60
the effective electrical impedance as
50
seen at the active sensor terminals. To
40
illustrate the E/M impedance method
30
we will consider experiments
performed on a spacecraft test panels
20
fabricated by NextGen Aeronautics,
10
Inc (Figure 6). The panel consisted of
0
a face skin, two I-beams, and four L150
200
250
300
350
400
450
500
550
600
650
700
Frequency (kHz)
shape stiffeners. The stiffeners were
bonded to the aluminum skin using a Figure 7 EM Impedance method: resonant
structural adhesive, Hysol EA 9394. frequencies spectrum showing increased amplitude
Damages were artificially introduced for the signal received at the sensor located on the
in the two specimens including cracks top of disbond DB1 (PWAS a2)
(CK), corrosions (CR), disbonds
ReZ
PWAS a2
PWAS a3
PWAS a1
(DB), and cracks under bolts (CB). The instrumentation set-up is presented in Figure 6b.
Among other seeded defects, the panel contained disbonds between the stiffeners and the skin.
They are of two types: partial disbonds DB1 and DB3, and through disbonds DB2 and DB4.
The corrosions are simulated as machined areas. The electromechanical impedance method was
used to detect disbonds, cracks and corrosions. It can be seen in Figure 7 that the resonant
spectrums of the signals from PWAS a1 and a3 located on an area with good bond are almost
identical. The resonant spectrum from PWAS a2 located on the disbond DB1 is very different
showing new strong resonant peaks associated with the presence of the disbond
ACKNOWLEDGEMENTS
The financial support of National Science Foundation award # CMS 0408578, Dr. Shih Chi Liu,
program director, Air Force Office of Scientific Research grant # FA9550-04-0085, Capt. Clark
Allred, PhD, program manager; and NASA STTR program Phase I topic T7-02 through
NextGen Aeronautics, Inc. are gratefully acknowledged.
REFERENCES
Bottai, G.; Giurgiutiu, V. (2005) “Simulation of the Lamb Wave Interaction between Piezoelectric Wafer
Active Sensors and Host Structure”, SPIE Sensors and Smart Structures Technologies for Civil,
Mechanical, and Aerospace Systems Conference, 6-10 March 2005, San Diego, CA, # 5765-29
Cuc, A.; Giurgiutiu, V. Tidwell, Z.; Joshi, S. (2005) “Non-destructive evaluation (NDE) of space
application panels using piezoelectric wafer active sensors”, Proceedings of the ASME IMECE
Congress, Orlando, FL, Nov. 7-11, 2005, paper #IMECE2005-81721, CD-ROM
Giurgiutiu, V. (2003) “Lamb Wave Generation with Piezoelectric Wafer Active Sensors for Structural
Health Monitoring”, SPIE vol. 5056, 2003, San Diego, CA, paper # 5056-17
Giurgiutiu, V. (2005) “Tuned Lamb-Wave Excitation and Detection with Piezoelectric Wafer Active
Sensors for Structural Health Monitoring”, Journal of Intelligent Material Systems and Structures,
Sage Pub., Vol. 16, No. 4, pp. 291-306, April 2005
Giurgiutiu, V.; Bao, J. (2004) “Embedded-Ultrasonics Structural Radar for In-Situ Structural Health
Monitoring of Thin-Wall Structures”, Structural Health Monitoring – an International Journal, Vol.
3, Number 2, June 2004, pp. 121-140
Giurgiutiu, V.; Cuc, A. (2005) “Embedded Nondestructive Evaluation for Structural Health Monitoring,
Damage Detection, and Failure Prevention”, Shock and Vibration Digest, Sage Pub., Vol. 37, No. 2,
pp. 83-105, March 2005
Giurgiutiu, V.; Jenkins, C.; Cuc, A. (2005) “New Results in the Use of Piezoelectric Wafer Active
Sensors for Structural Health Monitoring of Aerospace Vehicle Parts” Aeromat-2005, Integrated
Systems Health Monitoring Session, 6-9 June 2005, Orlando, FL (oral presentation only)
Raghavan A., Cesnik C. (2004) "Modeling of piezoelectric-based Lamb-wave generation and sensing for
structural health monitoring"; Proceedings of SPIE - Volume 5391, July 2004, pp. 419-430
Yu, L.; Giurgiutiu, V. (2004) “Advanced Signal Processing Techniques for Multi-Damage Detection
with an Improved Embedded Ultrasonics Structural Radar Algorithm and Piezoelectric Wafer Active
Sensors”, 2004 ASME Congress, Nov. 13-19, 2004, Anaheim, CA, paper # IMECE2004-60969
Yu, L.; Giurgiutiu, V. (2005) “Multi-damage Detection with Embedded Ultrasonic Structural Radar
Algorithm using Piezoelectric Wafer Active Sensors through Advanced Signal Processing”, SPIE
Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems
Conference, 6-10 March 2005, San Diego, CA, paper # 5768-48