INTEGRITY OF PIEZO-COMPOSITE BEAMS UNDER HIGH CYCLIC
ELECTRO-MECHANICAL LOADS - EXPERIMENTAL RESULTS
Lucy Edery-Azulay and Haim Abramovich
Faculty of Aerospace Engineering, Technion, I.I.T., Haifa 32000, Israel
[email protected]; [email protected];
ABSTRACT
One key issue in the study of the structural integrity of smart structures is the research of high cyclic
electro-mechanical (E/M) loading. To truly understand the integrity of smart structures, investigation
must include active sensors/actuators embedded or bonded as a part of a smart structure, and to be
loaded in a combined electro-mechanical cyclic loads as the way they will be used in real life
applications. A better understanding of the effects of cyclic E/M loading is necessary to characterize the
lifetime behavior of active ceramic components.
The present research is a part of a comprehensive1 experimental study dealing with the structural
integrity of smart structures. A laminate composite made of graphite-epoxy with a quasi-isotropic lay-up
was used as a host structure for all the specimens. Two commercially available piezoelectric patches
(PZT-5H, lead zirconate titanate) were used: ACX patches from QP15N Cymer Inc. U.S.A. and PIC255
from PI Ceramics, Germany. The first part of the present study investigates the behaviour of piezolaminated beams with embedded or surface bonded patches, subjected to axial tension/compression
loads. Seven different specimens (numbered as AX-EM) with embedded piezoceramic patches and
another two specimens (numbered as AX-SM) with four piezoceramic patches bonded to the surface of
each host structure (one pair of ACX and one pair of PIC-255 patches) were tested. The second part of
the present study investigates the behaviour of seven specimens (numbered as BEN1-7) with four
piezoelectric patches, one pair of ACX and one pair of PIC-255, bonded on each side of the host
structure. These specimens were subjected to bending cyclic loads using four-point bending test setup.
The degradation in the sensing abilities of the piezoelectric patches for an increasing number of E/M
load cycles was monitored and recorded.
The aim of the present research is to examine the changes in the piezoceramic sensing capabilities as
a piezo-laminated structure is undergoing an increasing number of electro-mechanical load cycles.
These two types of integrated loads, called Electro-Mechanical (E/M) loads can be applied in-phase
(both loads cause either tension or compression) or out-of-phase (one load causes tension while the
other causes compression and vice versa).
It is expected that the main impact of the present research will be its capability to predict the structural
integrity of a given smart structure as a function of its use, yielding a balanced design with an enhanced
survivability and a higher confidence in the usage of piezoelectric patches.
1
Yet, in general, structural life prediction involves analyzing data gathered from a large number of tests. The present review
does not pretend to cover the whole spectrum of experiment that have to be done.
INTRODUCTION
The key issue in the structural integrity of smart structures used for aerospace industry is its applied
cyclic loading. Therefore, to truly understanding the integrity of smart structures, investigation must
include active sensors/actuators embedded or bonded as a part of a smart structure, and to be loaded
in a combined electro-mechanical cyclic loads as the way they will be used in real-life applications.
Surveying the published experimental studies one can found only a few relevant experimental reports
where the life span of the piezoceramic patches is studied in connection to two types of stress
conditions: tension/compression due to axial loads and bending stresses. Bronowicki et al. [3]
performed one of the first studies on the structural integrity of a smart structure with embedded active
sensor/actuator. Graphite/epoxy laminates with embedded active PZT were loaded with various
external tensile mechanical loads, and then completely unloaded. PZT’s changes and degradation in its
performance were monitored. Paget et al. [4] evaluated the performance of embedded piezoceramic
transducers used as Lamb-wave2 generators. Composite specimens were subjected to static tensile
and compressive loads as well as fatigue loading. Mall and Coleman [5] investigated the effect of
embedded active PZT sensor/actuator on tensile strength and fatigue behavior of quasi-isotropic
graphite/epoxy laminates. A commercially PZT (ACX QP15N)3 was inserted in the two middle 90° plies.
One of the first studies that tested the influence of combined electric and mechanical loads, (E/M
fatigue) was presented by Mall and Hsu [6]. The behavior of quasi-isotropic graphite/epoxy laminates
with embedded PZT was studied. Only tensile mechanical loads were applied while the embedded PZT
was subjected to positive or negative AC voltage, giving either an in-phase or out-of-phase. Yocum and
Abramovich [2] presented another pioneering study, in which they investigated the functionality of an
embedded piezoceramic actuator/sensor undergoing fully reversed tension compression E/M fatigue.
The tested specimens were similar to those used in Ref. [6] and in the present study. To date, only a
few studies dealt with the life span of piezoelectric sensor/actuators under bending stress conditions,
although the integrity of smart structures must refer also to its behavior in the presence of bending
stress conditions. Mitrovic et al [7], investigated the behavior of piezoelectric materials under combined
electro-mechanical loading. Experimental results indicate strong dependence of both elastic and
piezoelectric properties on the operating conditions (mechanical and electrical). Yet, only static loads
were applied. Thielicke et al [8] investigated the reliability of the piezoelectric patches themselves and
their adhesion on the substrate. PZT patches were subjected to cyclic mechanical loads in a four point
bending system. Jones et al [9] characterized the strain accumulation of a ferroelectric/ferroelastic
ceramic PZT patch under mechanical cycling using four-point bending bar geometry. The present
research is a part of a comprehensive1 experimental study dealing with the structural integrity of smart
structures. The first part of the present research is a direct continuance of Ref [2]. It extends the
investigation of piezo-laminated beams under uniaxial electro-mechanical cyclic loads. The surface
mounted piezo-composite specimens were loaded with asymmetric E/M loads profile ranged from
100% to 150% limit stresses in in-phase conditions. The behavior of two commercial PZT patches, ACX
and PIC-255, was studied. Degradation in the sensing abilities of the piezoelectric patches was
monitored for an increasing number of load cycles. It was found that an out-of-phase tensioncompression (T-C) E/M fatigue caused larger damage to the PZT than in-phase T-C E/M fatigue for a
given applied maximum stress. Yet, far less damage was experienced by the PZT in the T-C tests than
in T-T tests (as compared with Ref. [6]). A stable output was found for the surface mounted PZT
patches loaded by in-phase asymmetric T-C, even when the applied loads were 150% of the limit
stress. The second part of the present research investigates the behavior of a lamina with a surface
mounted PZT subjected to bending cyclic loads. The specimens were mounted in a four point bending
fixture. Each specimen contained two types of commercial piezoelectric patches, ACX and PIC255. The
specimens were exposed to different load types: combined symmetric electro-mechanical cyclic loads,
2
Lamb wave generation is a way of determining damage in a structure. The system of embedded PZT's, as a Lamb wave
generator, would be part of a built-in structural health monitoring system.
3
Similar patches were used in the present research
in- or out-of-phase, only electric loads and only mechanical loads. Experimental results are presented
and discussed with respect to the results of the first part of the present research, the uniaxial
tension/compression loads, Ref [2, 6]. A good correlation was found. In addition, the application of only
tensile electric load cycles was found to have a significant destructive effect on the patch life span.
It is expected that the main impact of the present research will be its capability to predict the structural
integrity of a given smart structure as a function of its use, yielding a balanced design with an enhanced
survivability and a higher confidence in the usage of piezoelectric patches.
Experimental system and equipment
A 50 tons MTS servo-hydraulic test machine was used to apply a monotonic tensile/compression cyclic
load. A KEPCO 500 power supply/ amplifier together with an IEC F31 Function generator (to alter the
frequency) were used to supply an AC electric load to the specimens. A computer code uploaded on a
VISHAY '6000 system' machine was used to store strains, applied loads and output sensing voltages.
The applied load and the PZT sensing voltage are read in parallel on two oscilloscopes: an AGILENT
S4624A oscilloscope and a TEKTRONIX 2430A oscilloscope. A hand held DVM was used to measure
the capacity and resistance of the patches. All test instructions are input through a special program into
a computer which then controls the MTS machine. The test setup is presented in Fig. 1a, while the
tested specimens can be either axially loaded (Fig. 1b) or under four point bending loads (Fig. 1c).
(b)
(c)
(a)
Figure 1 – (a) The test setup (b) A specimen with buckling guide (c) Four point bending test setup
The specimens
A laminated composite beam made of graphite-epoxy with a quasi-isotropic lay-up4 was used as a host
structure for all the specimens. Two commercially available piezoelectric patches (PZT-5H, lead
zirconate titanate) were used: ACX patches from QP15N Cymer Inc. U.S.A. and PIC255 from PI
Ceramics, Germany. The ACX patches are pre-packaged with an outer protective polyimide5 layer and
extended lead wires. The PIC-255 patches have exposed silver electrodes and the wires (the terminals)
have to be soldered manually by the user. Table I summarizes some of the main material properties of
these two piezoceramic materials.
The embedded specimens group was characterized by a sole ACX piezoceramic patch being inserted
in the middle plane of the laminated composite beam (see Fig.2a), while the surface mounted
specimens had four piezoelectric patches bonded externally: one pair of ACX patches and one pair of
PIC-255 ones. Each pair was glued on both sides of the laminated composite beam. Back-to-back
strain gauges were placed at the mid-span of all the specimens, and were referred as S.G. 11-12.
Some additional strain gauges were used depending on the case load (see Fig.2b).
The axially loaded specimens were clamped from both sides. As our tests were to be tensioncompression tests, we used a buckling guide in order to prevent the possibility of buckling6, (see
Fig.1b). For the bending experiments, the specimens were mounted in a four-point bending fixture with
inner and outer spans of 170 and 200 mm. This arrangement yields a constant tension/compression
strain region located between the inner two reactions (see Fig.1c and Fig. 2d). The PZT patches (one
pair or more) were glued on each side of the host structure in this constant strain region. The specimen
lay in a stationary system, and was subjected to compressive mechanical loads applied by the MTS
machine. Figs. 2(a-d) illustrate schematically both the embedded and surface bonded specimens and
the four-point bending test setup.
Electro-mechanical cyclic loads
Changes in piezoceramic sensing capabilities were examined while a piezo-laminated structure was
undergoing an increasing number of electro-mechanical load cycles. E/M loads can be applied in two
different combinations, an in-phase or an out-of-phase condition. In-phase conditions refer to the case
where both the mechanical and the electric loads would induce on the patch and the specimen a strain
with equal tendency (either tension or compression). The out-of-phase condition refers to the case were
the two types of loads would induce on the PZT patch and the specimen7 opposite strain tendencies.
Both the electric and the mechanical loads were applied at a frequency of 10 Hz. Using a free strain
test analysis with frequencies between 1-10 Hz, the 10 Hz frequency was found to have the least
hysteresis for both piezoceramic actuators, ACX and PIC-255.
4
A composite laminate made of graphite-epoxy with [0/±45/90]S and [0/±45/90]3S was used as a host structure for axial and
bending test, respectively.
5
The protective polyimide improves the handing of the brittle PZT material while providing mechanical stabilization and
electrical insulation.
6
For more details about the complicated process of manufacturing these specimens and the design of the buckling guide see
Ref. [1].
7
The application of a negative electric field on a piezoceramic material whose piezoelectric coefficient d31 has a negative
value, would induce a tension strain, ε = d 31 E , where ε is the induced strain and E is the applied electric field.
Table I – Piezoelectric patches material properties
Unit
ACX
T
ε11 /ε0
Relative dielectric permittivity
Density
Piezoelectric charge coefficient
1800
7700
3
kg/m
d33
d31
Elastic modulus
E11
E33
Maximum suffer volt
Size (l ⋅b ⋅ t)
8
Calculated Capacity
m/V
or
C/N
N/m
304.8 mm
228.6 mm
Embedded
ACX PZT
400⋅10-
12
12
-180⋅10-
12
10
6.2⋅10
10
10
4.8⋅10
±400
46 x 26 x 0.2
75
6.9⋅10
10
5.5⋅10
±190
46 x 26 x 0.13
116
S.G.-14
S.G. -11-12
12
-179⋅10-
Volt
mm
ηF
1.03mm
1650
7800
350⋅10-
2
PIC-255
ACX-1
S.G.-11
ACX-2
S.G.-12
PI-3
S.G.-13
PI-4
Tab
50.8
mm
50.8
mm
50.8
PZT
wire leads mm
(a)
(b)
Front side
Back side
L=200 mm
ACX-1
P(t)/2
P(t)/2
ACX-2
S.G- 12
S.G.- 11
PI-3
PI-4
15
mm
P/2
PZT___
patches
15
mm
P/2
Graphite-epoxy
lamina
(c)
(d)
Figure 2 Axially loaded specimens: (a). Schematic embedded piezoelectric patches, (b). Schematic
surface bonded piezoelectric patches Bending experiments: (c) A typical schematic bending specimen
(d) A schematic four-point bending setup test
8
The patch capacity was calculated as follows: C = k ⋅ ε 0 ⋅ S = k ⋅ ε 0 ⋅ l ⋅ b ,
h
t
ε 0 = 8.85 ⋅ 10 −12
Test procedure
At the onset of each test (axial or bending) the polarity of the PZT patch must be known. Once the
polarity is determined, in-phase or out-of-phase conditions can be imposed. In addition, the capacity of
each piezoelectric patch at the beginning of the test and from time to time during the test is measured
using the DVM device9. The loading conditions were specifically determined for each specimen. As a
reference for the mechanical induced strains and the electrical load (voltage) we have used the limit
values suggested by the ACX company, 1000µε and 100 Volt, respectively10. Prior starting the test, the
initial output voltage for each PZT, Vout1, was measured. This voltage, Vout1, became an important value
to be monitored and recorded as it served as an index for the piezoceramic patch health. The sensing
voltage values were recorded peak-to-peak.
During the E/M cyclic load tests the cycling was stopped (at least once every half decade) to determine
the leftover PZT’s sensing capabilities, at the same loading conditions as Vout1 was found. If the
measured voltage was greater than 50% of Vout1, then the test was continued, otherwise, the PZT patch
was short-circuited and repoled. The repoling process was performed by applying a voltage of 250 DC
[V] to the patch for 7-15 minutes, for the embedded and surface mounted specimens, respectively.
After repoling, the output voltage is checked again. If the new output voltage is at least 70% of Vout1
then the test is continued, otherwise the PZT patch is considered to have failed.
Tests results and discussion
A test matrix of the completed tensile/compression and bending E/M tests is shown in Table II; Seven
different specimens with one embedded ACX piezoceramic patches (See Fig. 2(a)), designated as AXEM#11, were investigated under various tensile/compression symmetric E/M loads combinations. Two
specimens with two pairs of ACX and PI-255 piezoceramic patches bonded to the surface of each host
structure, (See Fig. 2(b)), nick-named as AX-SM#12, were subjected to a series of tests in asymmetric
tensile/compression. Seven different specimens with two pairs of piezoceramic patches (See Fig. 2(c)),
designated as BEN#13, were loaded in bending.
In general, during the bending tests, the PI patches would yield almost twice the initial sensing voltage,
as compared to the ACX patches, namely 110.0 and 60.0 volts, respectively. This can be explained by
the different piezoceramic material properties and thicknesses of the two patches. Yet, this difference
became much smaller during the axial tests, which yielded significantly lower sensing values (only a
few tens of volts) as compared to those monitored in the bending test. This is due to the different
contributions of axial and bending deformations in the sensing models, Ref [10].
According to the detailed test procedure, from time to time repoling process was necessary. In some
cases repoling enabled the patch to regain back almost all of its initial sensing voltage. Yet, during or
even from the beginning of some experiments, contact and/or short circuit problems caused the
"loosing" of different patches.
9
It was found that the measured value for the patch capacity, not necessary indicates the patch condition during the test.
No specific limit values were provided by the PI company
11
AX-EM#, AX-axial, EM-embedded patch, # - specimen number
12
AX-SM#, AX-axial, SM-surface mounted patch, # - specimen number
13
BEN#- BEN-Bending, #- specimen number
10
Bending experiments
Axially loaded experiments
Specimens
AX-EM 1
AX-EM 2
AX-EM 3
AX-EM 4
AX-EM 5
AX-EM 6
AX-EM 7
ASY#1
ASY#2
ASY#3
AX-SM 1
SY#4
SY#5
SY#6
ASY#1
AX-SM 2
SY#2
ASY#3
BEN1
BEN2
BEN3
BEN4
BEN5
BEN6
BEN7
Table II – Experimental test matrix
Mechanical
strain [µε]
Test type
Recorded on
S.G. 11-12
±500
In-phase
±1000
In-phase
±1500
In-phase
±1750
In-phase
±500
Out-of-phase
±1000
Out-of-phase
±1750
Out-of-phase
+500/-1000
+625/-1250
+750/-1500
In-phase
+1000/-1000
+1250/-1250
+1500/-1500
+500/-1000
+1000/-1000
In-phase
+1500/-1000
±1000
In-phase
±1000
In-phase
Out-of-phase
±1000
Out-of-phase
±1000
Electrical load
±1000
only
Electrical load
±1000
only
Mechanical
±1000
Load only
Electric
load
[Volt]
±100
±100
±100
±100
±100
±100
±100
+50/-100
+62.5/-125
+75/-150
+100/-100
+125/-125
+150/-150
+50/-100
+50/-100
+50/-100
±100
±100
±100
±100
±100
Ranged or
Total number
of cycles
6
0-1⋅10
6
0-1⋅10
0-1⋅106
6
0-1⋅10
6
0-1⋅10
6
0-1⋅10
6
0-1⋅10
6
0-2⋅10
6
6
2⋅10 -2.5⋅10
6
6
2.5⋅10 -3.5⋅10
6
6
3.5⋅10 -4.1⋅10
6
6
4.1⋅10 -4.3⋅10
6
6
4.3⋅10 -4.7⋅10
6
0-1.5⋅10
6
6
1.5⋅10 -2.0⋅10
6
2.0⋅10 -2.1⋅106
6
0-2.65⋅10
6
0-2.65⋅10
6
0-2 ⋅10
6
0-1.65 ⋅10
6
0-2 ⋅10
6
±100
0-0.5 ⋅10
±100
0-2.7 ⋅10
6
Figs. 3(a-b) shows typical normalized14 sensing output voltage of the piezoceramic patches as function
of the bending load cycles for specimens subjected to in-phase and out-of-phase conditions,
respectively. Referring to the in-phase results, an almost stable sensing response during more then two
million load cycles was encountered for the two commercial patches. On the other hand, much less
stable behaviour was monitored for the patches that were tested in out-of-phase conditions. It is
assumed that the in-phase condition provides a “stress relieving” effect which lowered the local
stresses on the PZT patch and thus preventing its damage. The out-of-phase condition increased the
local stresses on the patch, yielding a larger and faster accumulated damage, (more details in Ref. [1]).
A good correlation was found with the test results obtained for specimens subjected to
tensile/compression in phase E/M loads. Figs. 4(a-b) show the comparisons between the in-phase and
out-of-phase tension-compression symmetric E/M load cycles tests at applied stresses of 100% and
175% of the manufacture's suggested limit, respectively. Fig. 4(b) also includes the results of one test
that was performed under mechanical loads only, at an applied maximum stress of 175% of the
manufacturer’s recommended strain limit of the PZT. It is clear, that the manufacturer’s recommended
strain limit of ±1000ηε together with ±100 Volt for axial and bending functioning, is a good limit for both
patches, ACX and PIC-255.
Concerning the bending tests, the application of compression loads yielded a special case were the two
patches (ACX1 & PI3) bonded on the upper surface of the beam were subjected to tension mechanical
loads only while the two lower patches (ACX2 & PI4) were subjected to compression mechanical loads
only. Figs. 5(a-b) present the lifetime behaviour of typical upper (ACX1) and lower (ACX2) patch,
14
All data was normalized by the initial sensing voltage, Vout1.
respectively15. It is clear from both graphs that application of only tensile electric load cycles has a
significant destructive effect on the patch health. In this case, the patch is loaded with only negative
(against the patch poling direction) electric loads and never gets a repoling regime from the applied
external voltage.
1.20
1.00
0.80
0.60
0.40
0.20
ACX1-BEN1
PI3-BEN1
ACX1-BEN2
ACX2-BEN2
PI4-BEN1
PI4-BEN2
Load cycles
0.00
0
50000
100000
150000
200000
Normalized sensing voltage
Normalized sensing voltage
1.40
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
ACX2-BEN4
PI3-BEN4
PI4-BEN4
ACX1-BEN3
PI4-BEN3
Load cycles
0
250000
ACX1-BEN4
500000
1000000
1500000
2000000
(a)
(b)
Figure 3 Lifetime behaviour of typical patches in different E/M bending loads; (a) In-phase (b) Out of phase
Normalized sensing voltage
Normalized sensing voltage
1
0.9
In phase
0.8
Out of phase
Load Cycles
0.7
1
10
100
1000
0.9
0.7
In phase
Out of phase
Mechanical
0.5
Load Cycles
0.3
10000 100000 1E+06
1
10
100
1000
10000 100000 1E+06
(a)
(b)
Figure 4 Symmetric Tension-Compression E/M cyclic load, in-phase and out-of-phase results at
manufacturer's suggested limit loads (a) 100% (b) 175%
ACX1-BEN1
ACX1-BEN7
ACX1-BEN6
1.40
1.20
ACX1-BEN2
ACX1-BEN5
ACX1-BEN4
1.00
0.80
0.60
0.40
0.20
Load cycles
0.00
0
50000
100000
150000
ACX2-BEN2
ACX2-BEN5
ACX2-BEN3
1.40
Normalized sensing voltage
Normalized sensing voltage
1.60
1.20
ACX2-BEN7
ACX2-BEN6
1.00
0.80
0.60
0.40
0.20
Load cycles
0.00
200000
250000
0
50000
100000
150000
200000
(a)
(b)
Figure 5 The lifetime behaviour of typical patches in different E/M bending loads
(a) Patch bonded to the upper surface of the beam (b) Patch bonded to the lower surface of the beam
15
We choose to present results for patches from the same company, ACX.
250000
CONCLUSIONS
The present study investigated the functionality and fatigue response of a piezoceramic actuator/sensor
(PZT) embedded or bonded on the surface of a laminated composite beam. Two commercially
available PZT's were used: ACX patches from QP15N Cymer Inc. and PIC255 from PI Company. A
graphite/epoxy laminate with a quasi-isotropic lay-up was used as a host structure. Different
combinations of electromechanical (E/M) loading cycles profiles were applied on each specimen,
supplying either in or out of phase conditions.
Combined tension-compression E/M loading cycles with symmetric and asymmetric profiles were
applied on the specimens according to the limit stresses (strains) suggested by the manufacturers,
(1000µε together with 100 Volts). In addition, tests with bending E/M cyclic loads using a four-point
bending setup were performed. The degradation in the sensing abilities of the piezoelectric patches for
an increasing number of E/M load cycles was monitored and recorded. The effect of different E/M loads
combinations on the sensing patch health was investigated. All results were analyzed and a
comprehensive comparison was performed between the present results (Ref. [1]) and the results
published in literature [6].
It was found that for a given electrical load, the increasing of the mechanical load will raise the sensing
degradation. Under equal direction electro-mechanical loads, namely the in-phase condition, leads to
only a minor degradation of the PZT patch while for a out-of-phase condition the damage is higher and
the degradation appears earlier. For a given mechanical load applying a positive electric load will
improve the patch’s sensing capabilities as compared to a negative applied one. The application of only
an electrical load, which would induce tensile stresses on the PZT patch, has a significant destructive
effect on the patch health.
It was also found that asymmetric E/M loads, where the compression load is larger then the tension
one, would improve the piezoceramic patch health. In contrast, application of asymmetric load where
the tension load is larger then the compression one would lead to a rapid degradation of the PZT Patch.
ACKNOWLEDGEMENTS
The authors would like to acknowledge the outstanding experimental work performed by Mr. A. Grunwald and his exceptional
assistance in setting the tests and dedicated assistance in performing them, and the assistance of Mrs. R. Yaffe, both from the
Aerospace Structures Laboratory, Faculty of Aerospace Engineering, Technion, Haifa, Israel
REFERENCES
1.
Edery-Azulay, L. and Abramovich, "Piezo-Composite Beams under High Cyclic Electro-Mechanical Loads - Experimental
Results", TAE Report No.971 Haifa, Israel: Faculty of Aerospace Engineering, Technion, IIT; January 2007.
2. Yocum, M., Abramovich, H., Grunwald, A. and Mall, S., “Fully Reversed
Electromechanical Fatigue Behavior of
Composite Laminate with Embedded Piezoelectric Actuator/sensor”, Smart Materials and Structures, 12 (2003), 1-9.
3. Bronowicki, A. J., McIntyre, L. J., Betros, R. S., and Dvorksy, G. R., “Mechanical Validation of Smart Structures” , Smart
Materials & Structures, Vol. 5, 1996, 129-139.
4. Paget, C. A. and Levin, K., “Structural Integrity of Composites with Embedded Piezoelectric Ceramic Transducer” , SPIE
Conference on Smart Structures and Integrated Systems (Newport Beach, CA, March), SPIE Vol. 3668, 1999, 306-313.
5. Mall, S. and Coleman, J. M., “Monotonic and Fatigue Loading Behavior of Quasi-Isotropic Graphite/Epoxy Laminate
Embedded with Piezoelectric Sensor”, Smart Materials & Structures, Vol. 7, 1998, 822-832.
6. Mall, S. and Hsu, T. L., “Electromechanial Fatigue Behavior of Graphite/Epoxy Laminate Embedded with Piezoelectric
Actuator”, Smart Materials & Structures, Vol. 9, 2000, 78-84.
7. Mitrovic, M., Carman, G. P. and Straub, F. K., "Response of piezoelectric stack actuators under combined electromechanical loading", International Journal of Solids and Structures, Volume 38, Issues 24-25, Pages 4357-4374.
8. Thielicke, B., Gesang, T., and Wierach, P., "Reliability of piezoceramic patch sensors under cyclic mechanical loading",
Smart Mater. Struct. 12 993–996, 2003.
9. Jones, J. L., Salz, C. R. J. and Hoffman, M., " Ferroelastic Fatigue of a Soft PZT Ceramic", Journal of the American
Ceramic Society, 88 [10] 2788–2792, 2005
10. Edery-Azulay, L. and Abramovich, H., Actuation and sensing of shear type piezoelectric patches-closed form solutions,
Composite Structures, 64, pp. 443-453, 2004.