Effects of Elevated Temperature on Guided-Wave

8/6/2019 Effects of Elevated Temperature on Guided-Wave

1/17

http://jim.sagepub.com/

StructuresJournal of Intelligent Material Systems and

http://jim.sagepub.com/content/19/12/1383The online version of this article can be found at:

DOI: 10.1177/1045389X07086691

2008 19: 1383 originally published online 20 May 2008Journal of Intelligent Material Systems and StructuresAjay Raghavan and Carlos E.S. Cesnik

Effects of Elevated Temperature on Guided-wave Structural Health Monitoring

Published by:

http://www.sagepublications.com

can be found at:Journal of Intelligent Material Systems and StructuresAdditional services and information for

http://jim.sagepub.com/cgi/alertsEmail Alerts:

http://jim.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://jim.sagepub.com/content/19/12/1383.refs.htmlCitations:

by guest on October 5, 2010jim.sagepub.comDownloaded from
http://jim.sagepub.com/http://jim.sagepub.com/http://jim.sagepub.com/http://jim.sagepub.com/content/19/12/1383http://jim.sagepub.com/content/19/12/1383http://www.sagepublications.com/http://jim.sagepub.com/cgi/alertshttp://jim.sagepub.com/cgi/alertshttp://jim.sagepub.com/subscriptionshttp://jim.sagepub.com/subscriptionshttp://jim.sagepub.com/subscriptionshttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsPermissions.navhttp://jim.sagepub.com/content/19/12/1383.refs.htmlhttp://jim.sagepub.com/content/19/12/1383.refs.htmlhttp://jim.sagepub.com/http://jim.sagepub.com/http://jim.sagepub.com/http://jim.sagepub.com/http://jim.sagepub.com/content/19/12/1383.refs.htmlhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsReprints.navhttp://jim.sagepub.com/subscriptionshttp://jim.sagepub.com/cgi/alertshttp://www.sagepublications.com/http://jim.sagepub.com/content/19/12/1383http://jim.sagepub.com/


2/17

Effects of Elevated Temperature on Guided-waveStructural Health Monitoring

AJAY RAGHAVAN AND CARLOS E. S. CESNIK*

Department of Aerospace Engineering, The University of Michigan, 1320 Beal AvenueAnn Arbor, Michigan 48109, USA

ABSTRACT: Elevated temperatures can cause significant changes in guided-wave (GW)propagation and transduction for structural health monitoring (SHM). This work focuses onGW SHM using surface-bonded piezoelectric wafer transducers in metallic plates for thetemperature range encountered in internal spacecraft structures (201508C). First, studiesdone to determine a suitable bonding agent are documented. This is then used in controlledexperiments to examine changes in GW propagation and transduction using PZT-5Apiezoelectric wafers under quasi-statically varying temperature (also from 20 to 1508C).Modeling efforts to explain the experimentally observed increase in time-of-flight and changein sensor response peak-to-peak magnitude with increasing temperature are detailed.Finally, these results are used in detection and location of mild and moderate damage usingthe pulse-echo GW testing approach within the temperature range.

Key Words: structural health monitoring, damage prognosis, guided waves, Lamb waves,

thermal variation, temperature compensation, spacecraft structures, integrated systems health

management, bonding agent.

INTRODUCTION

Motivation

IN 2005, a new vision was defined for NASA, and it

has already set in motion plans for returning

astronauts to the Moon, and eventually, longer-term

missions to Mars. In this endeavor, integrated systemshealth management (ISHM) will play a key role in

fulfilling the mission objectives. ISHM will help in

transitioning from low-earth orbit missions with con-

tinuous ground support to more autonomous long-term

missions (Carlos et al., 2005). The ISHM system will

manage all the critical spacecraft functions and systems.

It will include monitoring capability composed of

sensors and actuators plus a reasoning system that can

evaluate the hardwares functionality. The structural

health monitoring (SHM) subsystem will be a crucial

component of the ISHM system. It will apprize

astronauts on changes in vehicle structural integrity

requiring action as well as providing the crew with thecapability to forecast potential problems and schedule

repairs based on the rate of loss of system function and

condition of hardware.

Fundamentals of Guided-wave Approaches

Among various technologies under investigation for

SHM, there are guided-wave (GW)-based approaches

These essentially involve exciting the structure with

high frequency GWs and processing the difference in

structural response with respect to a baseline signal for

the pristine condition using a tested algorithm to detectdamage and characterize it, if present. Guided waves can

be defined as stress waves forced to follow a path

defined by the material boundaries of the structure

For example, when a beam is excited at high frequency

stress waves travel in the beam along its axis away from

the excitation source, i.e., the beam guides the waves

along its axis. While several transducers have been

tested, piezoelectric wafer transducers (hereafter referred

to as piezos) seem to be the most commonly used

option. This is largely because of the low mass and space

penalty associated with incorporating them (crucial in

aerospace structures) and their high energy density for

high frequency applications. There are two mainapproaches commonly used in GW SHM, pulse-echo

and pitch-catch. In the former, typically after exciting

the structure with a pulse, a sensor (usually immediately

adjacent to the actuator) is used to detect scattered

echoes of the pulse coming from discontinuities. Since

the boundaries and the wavespeed for a given cente

actuation frequency of the toneburst are known*Author to whom correspondence should be addressed.E-mail: [email protected] 16 and 823 appear in color online: http://jim.sagepub.com

JOURNAL OF INTELLIGENT MATERIAL SYSTEMS AND STRUCTURES, Vol. 19December 2008 1383

1045-389X/08/12 138316 $10.00/0 DOI: 10.1177/1045389X07086691 SAGE Publications 2008

Los Angeles, London, New Delhi and Singapore

http://jim.sagepub.com/http://jim.sagepub.com/http://jim.sagepub.com/http://jim.sagepub.com/


3/17

the signals from the boundaries can be filtered out

(assuming a narrow enough time-span excitation pulse is

used; alternatively one could subtract the test signal

from the baseline signal). One is then left with signals

generated from the damage sites (if present). From these

signals, damage sites can be located using the known

wavespeed. One potential caveat of this approach isthat there are small areas in the immediate vicinity of

the actuator and structural features (e.g., stiffeners) with

decreased sensitivity for the actuatorsensor pair. The

actual size of these areas decreases with decreasing

time-span of the excitation pulse. This results from the

difficulty in distinguishing between minor differences in

the actuation signal/reflections from structural features

(e.g., caused by small temperature changes) and reflec-

tions caused by structural damage.

In the pitch-catch approach, a pulse signal is usually

sent across the specimen under interrogation and a

sensor at the other end of the specimen receives the

signal (in plate-like structures this is often implementedusing a network of such paths to cover the structure).

From various characteristics of the received signal, such

as delay in time-of-flight, amplitude, frequency content,

etc., information about the damage (if present) can be

obtained. Thus, the pitch-catch approach cannot be

used to locate the defect unless a dense network of

transducers is used. Issues with both approaches in the

face of changing temperature will be discussed later in

this study. A detailed survey of GW SHM, including

fundamentals and early history, is presented in a review

paper by the authors (Raghavan and Cesnik, 2007c). As

pointed out there, while GW SHM has shown a good

deal of promise in various laboratory demonstrations,several issues remain to be resolved before they see

widespread field deployment in structures. Among these,

compensation for environmental and service conditions

is a crucial one.

Spacecraft Structures and Their Environment

Typical spacecraft structures are composed of different

substructures, each of which can act as waveguides,

thereby making them attractive application areas for GW

SHM. This is also true of the planned NASAs crew

exploration vehicle (CEV). The CEV is expected to have

an aluminum alloy internal structure in the shape ofa blunt body capsule protected by bulk insulation,

composite skin panels, and a thermal protection system

(TPS), see Stanley et al. (2005). Spacecraft structures

in particular present a challenging application due to

the harsh environment of outer space as well as the

tremendous heat flux and high temperatures attained

during re-entry into a planets atmosphere. While fiber

optic sensors can be designed to withstand these harsh

re-entry temperatures, they are passive and cannot be

used to excite GWs. Some authors have examined SHM

in spacecraft structures using fiber optic sensor networks

by passive strain/loads monitoring (e.g., Friebele et al.,

1999; Ecke et al., 2001). However, such passive

approaches require a much denser network of sensors

than GW approaches. The internal spacecraft structures,

however, are somewhat insulated by the TPS. The TPS is

typically designed to keep temperatures below 1508C ininternal structures, particularly in manned missions

(Myers et al., 2000). Apart from the re-entry phase,

even in the course of the flight, the temperature of

spacecraft structures varies significantly, with tempera-

tures up to 708C (Larson and Wertz, 1995) depending on

whether they face towards or away from the Sun. For

solar arrays, this fluctuation is even greater (up to 1008C,

see Larson and Wertz, 1995). Another source of

temperature variation in internal spacecraft structures is

the heat radiated by cabin electronics, which is difficult to

reject into space, and is therefore controlled by active

cooling. Commercial piezos are functional without loss in

properties up to half their Curie temperature. For leadzirconium titanate, or PZT 5A (a.k.a. DoD Type II), one

of the more commonly used piezoceramics, half the Curie

limit is about 1758C. Thus, internal spacecraft structures

become a potential application area for GW SHM using

PZT-5A piezos. Some studies have examined GW SHM

for cryogenic tanks (Blaise and Chang, 2001; Lin et al.,

2003) and thermal protection systems (Yang et al., 2003;

Derriso et al., 2004; Huang et al., 2005) in spacecraft

structures. However, the experiments/simulations in

these studies were restricted to room temperature

(except in Blaise and Chang, 2001). The GW SHM

algorithm must account for temperature changes to

minimize false damage indications and reduce errors indamage characterization during the course of space

missions. The present study explores this very issue.

Previous Efforts on Effects of Temperature for GW SHM

There have been some efforts to address the issue of

varying temperature for GW SHM in the literature.

Blaise and Chang (2001) investigated the performance

of piezoelectric transducers (in GW pitch-catch config-

uration) embedded into sandwich structures for cryo-

genic fuel tanks at low temperatures (up to 908C).

An empirical model (linear) was fitted to experimentally

obtained data points for changing signal peak-to-peakmagnitude and time-of-flight. Reasonable agreement

between the interpolated signals from the empirical

model and experimental data for intermediate tempera-

ture values was obtained. However, no damage detec-

tion studies were reported by them. Lee et al. (2003)

studied the effect of temperature variation on the Lamb-

wave response of a piezoceramic sensor in a pitch-catch

configuration on a metallic plate from room tempera-

ture up to 708C. They observed that the effect of

temperature variation over this range (analyzed using

1384 A. RAGHAVAN AND C. E. S. CESNIK



4/17

principal component analysis) was much more

pronounced than the effect of damage (drilled 1-mm

diameter hole). Lu and Michaels (2005) and

Konstantinidis et al. (2006) examined GW SHM under

mild thermal variations from 20 to 408C. Both addressed

modeling the varying time-of-flight in this temperature

range due to changing substrate elastic modulus andthermal expansion, and good agreement with experi-

ments in this temperature range was observed. Lu and

Michaels (2005) also suggested using a bank of baseline

signals for various temperatures and picking a baseline

signal which minimizes difference relative to the test

signal for that particular temperature. Chambers et al.

(2006) suggested test procedures for environmental

robustness certification of GW SHM transducers in

aircraft structures. Schulz et al. (2003) studied the

performance of PZT-5A patches as free vibration

sensors bonded using various adhesives on aluminum

beams up to 2408C. A drop of 74% in strain response

amplitude (relative to room temperature value) wasobserved at 1508C, and the response dropped to zero at

2408C. They also explored alternative materials for high

temperature applications, and identified lithium niobate

and nanotubes as potential transducer materials for

future high temperature applications (up to 10008C).

However, these have much weaker piezoelectric proper-

ties at room temperatures compared to PZT. Results

from preliminary thermal experiments were also

reported by the authors (Raghavan and Cesnik,

2005a). These tests examined sensor response variation

in pitch-catch tests done using PZT-5A piezos bonded

on a 1-mm thick aluminum plate up to 1508C. It was

observed that the sensor response decreased continu-ously with increasing temperature and went to the noise

floor beyond 1108C. It was suspected that this decrease

was due to bonding agent degradation.

Objectives of this Study

It is evident from the literature reviewed that the

issues of compensation for and damage characterization

under thermal variations expected in GW SHM for

spacecraft structures (above room temperature) have

not received much attention. This study aims to

contribute in these aspects. First, studies done to find

a suitable bonding agent (for GW SHM using piezo-ceramics on aluminum plates) that does not degrade

under thermal variations from 20 to 1508C are reported.

With a suitable bonding agent chosen, controlled

experiments are done to examine changes in GW

propagation and transduction using PZT-5A piezos

under quasi-statically varying temperature (also from

20 to 1508C). All parameters changing with temperature

are identified and quantified based on data from

the literature. Modeling efforts exploiting these data

to explain the experimental results are outlined.

Finally, these results are used to explore detection

and location of damage (indentations/holes) using the

pulse-echo GW testing approach in the same tempera

ture range.

BONDING AGENT SELECTION

Bonding Agents Tested and Specimen Preparation

After an initial pre-screening, three different two-par

epoxies were evaluated for the temperature range o

interest. These were 10-3004 (Epoxies, Etc., 2006), and

Epotek 301 and 353ND (Epoxy Technology, 2006)

Epotek 301 and 353ND, both low-viscosity agents, are

rated for continuous operation up to 200 and 2508C

respectively. 10-3004 is relatively viscous, and is rated

for continuous operation up to 1258C, although the

manufacturer clarified that it should work up to 1508C

for short-term use (hours). In addition, it was confirmedfrom the manufacturers that each epoxy would be

suitable for surface-bonding piezoceramics (with metal-

lic electrodes) on aluminum plates. While 10-3004 and

301 can be cured overnight at room temperature

353ND needs to be cured in an oven at 808C fo

25 min. Standard surface preparation procedures were

followed with each, i.e., the plate surface was made

rough by light sanding, and both the plate and piezos

were cleaned thoroughly using acetone to get rid of

grease and dust. After uniformly applying a thin-layer

film of epoxy to both surfaces and cleaning the excess

light pressure was applied using small weights (2 lb.) to

the interface to help the bond set.

Experimental Tests with Epotek 301 and

Epoxies, Etc. 10-3004

The first aluminum alloy (5005) plate specimen

(40350.32cm3) tested had four PZT-5A piezo

that were surface-bonded using Epotek 301 (Figure 1)

Two piezos were used as actuators (dimensions

2.51.50.03 cm3, at the center, on either surface)

and two as sensors (dimensions 1 10.03 cm3). One o

the sensors (Sensor 1) was immediately adjacent to the

actuator, and the other (Sensor 2) was 10 cm away from

the plate center. This specimen was thermally cycledfrom 20 to 1508C in an industrial oven and then cooled

back to room temperature over three cycles. A Labview

based automated thermal test setup was developed for

these experiments. After turning the oven on, at every

108C intervals (read by a type-K thermocouple with

18C accuracy attached to one side of the plate

specimen), the Labview program triggered an Agilen

33220A function generator to send a 3.5-cycle Hanning-

windowed toneburst, with center frequency 210 kHz to

the actuators (excited symmetrically), 16 times each

Effects of Elevated Temperature on Guided-wave SHM 1385



5/17

at 1-s intervals. A Hewlett Packard 54831B Infiniium

oscilloscope recorded the sensor response signals, which

were sampled at 10 MHz and averaged over the 16

readings at each temperature. In these tests, it was

observed that the sensor response signal of Sensor 2decreased monotonically in peak-to-peak magnitude

with increasing temperature (Figure 2). The error bars

shown are based on the standard deviation over the 16

readings at each temperature. Furthermore, Sensor 2s

response signal peak-to-peak magnitude at room tem-

perature decreased at the end of each cycle, and the

shape of the signal also changed significantly, as shown

in Figure 3. It should be noted that the sensor response

was compensated for varying actuation signal magni-

tude (which dropped due to the increasing capacitance

of the actuators with temperature). While some amountof irrecoverable loss in response strength is expected

after the first few cycles, due to thermal pre-stabilization

of piezos (Berlincourt et al., 2007), the signal shape is

not expected to change. Despite the actuators being

excited symmetrically, and thereby only supposedly

exciting the S0 mode, experimental imperfections cause

weak A0 mode excitation. To counter this, the sensor

was originally designed (Raghavan and Cesnik, 2005b)

to be very weakly sensitive to the A0 mode (the sensor

size equaled the A0 mode wavelength at 210 kHz).

However, after each cycle, the strength of the slower A0mode contribution at 208C in the Sensor 2 signal

increased. This suggested that the sensors effectivearea kept decreasing after each cycle. Based on these

factors, it was concluded that the Epotek 301 bond line

was indeed degrading as a result of the thermal cycling.

An analogous test was done with PZT-5A transducers

bonded using 10-3004 on an aluminum alloy (5005)

plate. In this case, the results were even more drastic and

the sensor response dropped gradually to the noise floor

at 1008C while heating in the very first cycle, and never

recovered.

Experimental Tests with Epotek 353ND

Finally, tests were done with Epotek 353ND.The specimen tested (also an aluminum 5005 specimen

of size 4035 0.32 cm3) was similar to the ones tested

above. The schematic of this is shown in Figure 4. It had

two 2 10.03 cm3 piezos surface-bonded on either face

of the plate at the center which were used as actuators.

Three surface-bonded 1 10.03cm3 piezos were used

as sensors, of which one was immediately adjacent to one

of the actuators and two were at a distance of 10.2 cm

from the plate center. The specimen was thermally cycled

in the same temperature range seven times in the oven,

35cm PZT-5A actuators2.5cm x 1.5cm

10 cm

3.2 mm

PZT-5ASensor 2

PZT-5ASensor 1

1 cm x 1 cm1 cm x 1 cm

Aluminum 5005 plate

Type K thermocouple

0.3 mm

40 cm

Supportblocks

Figure 1. Schematic of specimen for tests with Epotek 301.

2 4 6 8 10

x105

0.12

0.1

0.08

0.06

0.04

0.02

0

0.02

0.04

0.06

Time (s)

Sensor

2signal(V)

Before thermal expts.

After thermal cycle 1



EMI

S0mode A0mode

Boundaryreflections

Figure 3. Sensor 2 signal at room temperature before and aftereach of the three thermal cycles for tests with Epotek 301 (thesignals are offset by a small DC voltage for clarity; EMI electromag-netic interference from the actuation).

0

0.05

0.1

0.15

0.2

0 20 40 60 80 100 120 140

Sensor

2peak-to-peakmagnitude(V) 1st cycle heating 1st cycle cooling

2nd cycle heating 2nd cycle cooling

3rd cycle heating 3rd cycle cooling

Temperature (deg C)

Noise floor

Figure 2. Variation of Sensor 2 response magnitude (peak-to-peak) and associated error bars with temperature over three thermalcycles (for tests with Epotek 301).




6/17

using the same experimental setup and settings described

above. In this case, the sensor response peak-to-peak

magnitude and shapes did not change (within negligible

error margins, see Figure 5) when the signals before and

after each thermal cycle are compared. The very firstcycle was an exception, being the thermal pre-stabiliza-

tion cycle discussed before, which caused a 17% drop in

Sensors 2 and 3 response peak-to-peak magnitude (but

the signal shape did not change after the cycle). Thus, this

epoxy proved to be suitable for the purposes of this study.

Thereafter, more controlled tests were conducted with

this same specimen to study signal changes for pristine

specimens and to explore damage characterization

at different temperatures, which is discussed in the

following sections.

EFFECTS OF ELEVATED TEMPERATURE

FOR PRISTINE SPECIMEN: EXPERIMENTS

AND MODELING

Experimental Setup and Results

In designing the specimen for tests with Epotek

353ND, Sensor 1 was collocated with the actuator

with the intention of using it for damage detection using

pulse-echo tests. Sensors 2 and 3 were for tracking

changes in the GW transmitted signal with temperature

(in undamaged state). In addition, Sensors 2 and 3

also act as mild GW scatterers, due to the increased

local stiffness and mass caused by their presence

This simulates the effect of some structural feature

(e.g., rivets) which could act as GW scatterers in morecomplex structures. While the specimen was tested in the

industrial oven initially to check for bond degradation

the ovens heating/cooling rate could not be tightly

controlled, and was very rapid (up to 108C/min) a

times. This fast heating rate led to non-repeatable

signals for Sensor 1, which could potentially be

interpreted as false positives. This is discussed in the

next section. More controlled tests were subsequently

done in a computer-controlled autoclave (Figure 6)

where both the heating and cooling rates were set to

18C/min. A 5-min dwell period at 1508C was also

included in the thermal cycle between the heating and

cooling phases (Figure 7). The data at 90 and 1008Cwhile cooling was not used, since in this temperature

range, the autoclave switches from exclusively air

cooling to a combination of air and water cooling

leading to oscillations in the cooling rate over this range

For these tests, the center frequency was reduced to

120 kHz. This was to minimize actuation signal distor

tion effects at higher frequencies caused by increasing

actuator capacitance at higher temperatures. While

a Krohn-Hite 7500 wideband amplifier was tried for a

couple of thermal cycles in the oven, it was unable to

35cm

Piezo actuators2 cm x 1 cm10.2

cm

3.2 mm

Aluminum 5005 plate

Sensor 31 cm x 1 cm

Thermocouple

Damage location

8 cm

0.3 mm

Sensor 2

1 cm x 1 cm

40 cm Supportblocks

1 cm x 1 cmSensor1

Figure 4. Schematic of specimen for tests with Epotek 353ND.(Damage introduced and discussed later.).

0 1

x104

0.05

0.04

0.03

0.02

0.01

0

0.01

0.02

0.03

0.04

0.05

Sensor2response(V)

Before thermal cycleAfter thermal cycle

S0 mode

Boundary reflections

Mild A0mode

Electromagneticinterference

Time (s)

0.2 0.4 0.6 0.8

Figure 5. GW signal sensed by Sensor 2 (bonded using Epotek3 53 ND ) bef or e a nd a fte r a th er m al c yc l e: th e dif f ic ul ty in distinguishing between the two signals is indicative of therepeatability of signals after the thermal cycle.

Autoclave

Plate specimenand cable stand

TC

Oscilloscope

Labview system

Function generator

Data acquisition for TC

Figure 6. Labeled photograph of setup and autoclave for controlledthermal experiments (TC thermocouple).




7/17

amplify without significant signal distortion and ripple

at higher temperatures. Therefore, no amplifier was usedfor the controlled tests in the autoclave. The actuation

signal was still a 3.5-cycle Hanning windowed toneburst

of 18 V peak-to-peak magnitude at 208C, with the two

actuators on either surface excited symmetrically. At

every 108C intervals, the LABVIEW setup triggered the

function generator to send this actuation signal to the

function generator 30 times with 1-s gaps between each

trigger. The 1-s gap between each excitation ensures that

the GWs from the previous excitation have died out and

do not interfere with the GWs from the next excitation

burst. The oscilloscope in turn is triggered by a square

wave pulse from the function generator through a

separate synchronization channel coinciding with thebeginning of each excitation burst. This ensures that

the start of the actuation pulse is set as time zero for

the signal in the oscilloscope. The oscilloscope stores the

signals from the piezos, averaged over the 30 signals

from the multiple excitation bursts to reduce signal noise

levels, along with the standard deviation in peak-to-peak

magnitude. During this 30-s period over which signal

averaging is done, the autoclave continues heating/

cooling at 18C/min, meaning the specimen temperature

varies by up to 0.58C. However, this temperature

variation is within the accuracy of the thermocouple

(18C). It is also representative of situations encoun-

tered in practical applications: one can seldom expectthe structure to remain at perfect thermal equilibrium

for on-the-field applications. Data were collected over

two thermal cycles for the pristine, undamaged condi-

tion. As mentioned before, as temperature increases, due

to increasing actuator capacitance, there is an increase in

the electrical load seen by the function generator.

Therefore, due to the lack of an amplifier, there was a

drop in actuation peak-to-peak magnitude from 18 V at

208C to %13V at 1508C (but negligible shape distor-

tion). All signals presented for higher temperatures in

this study have been scaled for 18-V peak-to-peakactuation level (by multiplying the sensor signals with

the ratio of 18V to the actuation peak-to-peak

magnitude at the corresponding temperature).

Figure 8 shows the GW signal read by Sensor 2 at

various temperatures while heating. Evidently, there is a

decrease in GW speed of the first transmitted GW pulse

as temperature increases. In addition, the signal peak-

to-peak magnitude increases with increasing tempera-

ture up to a certain point (around 908C) and then

decreases with increasing temperature. Hysteresis effects

were found to be negligible (unlike in the oven tests,

where significant hysteresis was observed between the

heating and cooling phases due to very differenttemperature change rates in the two phases).

Identification and Quantification of Thermally

Sensitive Parameters

In order to explain these effects, an effort was made to

identify all parameters in the experiment that change

with temperature. The following list was compiled and

data for their thermal variation were found from various

sources in the literature:

1. Youngs moduli of structural substrate and PZT-5A:

The substrate elastic modulus is probably the mostimportant parameter for thermal variations. There is

a significant decrease in the elastic modulus of

aluminum with increasing temperature. This causes

a reduction in GW speeds, as reflected in the change

in dispersion curves. Furthermore, in quantifying

thermal variations of elastic modulus, two different

data sets were found: one for the variation in static

elastic modulus (U.S. Munitions Board Aircraft

Committee, 1955), and the other for dynamic elastic

modulus (Lord and Orkney, 2000). These data are

0 2 4 6 8 10

x 105

0.06

0.04

0.02

0

0.02

0.04

Sensor2response(V)

20C90C (heating)150C

Time (s)

Figure 8. GW signals recorded by Sensor 2 (averaged over30 signals for each temperature) at various temperatures whileheating (offset by a small DC voltage for clarity).

0 1 2 3 4 50

50

100

150

Tempera

ture(C)

Time (h)

Figure 7. Typical timetemperature curve for experiments done inthe computer-controlled autoclave.




8/17

shown in Figure 9. The former was obtained from

standard stressstrain tests conducted under varying

temperature for aluminum alloy 7075, while the latter

was found from measuring changes in natural

frequency of aluminum beam (alloy 5052) flexural

vibrations with temperature. No data were found foraluminum 5005, the material used in the tests here.

However, it is similar in composition to the two

alloys for which data were found. The variation in

elastic modulus of PZT-5A is relatively small

(Williams et al., 2004). No data were obtained for

dynamic elastic modulus variation of PZT-5A.

2. Piezoelectric properties of PZT-5A: It is well-known

that the piezoelectric constants (d31 and g31) vary

significantly with temperature (Berlincourt et al.,

2007). For GW SHM, the variation in the product

d31 g31 is of relevance (the d31 constant is associated

with actuation shear stress induced, while the g31

constant is associated with the piezo-sensor sensitiv-ity), and this can vary by as much as 7%, as shown

in Figure 10. In addition, the dielectric constant of

PZT-5A increases linearly with temperature, which

causes the load seen by the function generator to

increase. This however, does not affect sensor

response by itself.

3. Thermal expansion: This is a relatively mild effect,

and causes the plate thickness, piezo dimensions, and

distances traveled by the GWs in the plate to

increase, while material density decreases. Since the

thermal expansion coefficients of aluminum and

PZT-5A are known (average values over 20 to

1508C are 25.5mm/m-8C for aluminum obtainedfrom Matweb (2007); 1 2.5mm/m-8C for PZT-5A,

see Williams et al., 2004), these effects can be

accounted for. The effect of changing (static) elastic

modulus, plate thickness, and density were used to

compute Lamb-wave phase velocity dispersion curves

at different temperatures (Figure 11). The dispersion

curves at equally spaced temperatures are unequally

spaced, particularly at the higher temperatures. This

is because of the nonlinear change in Youngs

modulus and density with temperature.

4. Damping and pyroelectric effects (not considered)

Another parameter that changes with increasing

temperature is damping in the structural substrate

The best reference found in this regard (Hilton and

Vail, 1993) estimated an increase by a factor of 4 in

the loss modulus (representative of damping) a

100 Hz in aluminum alloy 2024. This still ensure

that the loss modulus is orders of magnitude lower

than the elastic modulus and can be neglected, as was

verified at room temperature in a previous work

(Raghavan and Cesnik, 2005b). Therefore, damping

was ignored at higher temperatures too. Finally, due

to the pyroelectric effect, temperature changes causea static voltage to appear across a piezos electrodes

Since the experimental signals were acquired with a

2 Hz high-pass filter built into the oscilloscope, thi

effect was not considered either.

Theoretical Modeling and Signal Analysis

Effects (a)(c) were incorporated into the theoretica

models developed by the authors (Raghavan and

Cesnik, 2005b). These models can capture the

1540

1560

1580

1600

1620

1640

1660

1680

0 20 40 60 80 100 120 140 160

PZT-

5Ad31xg31

(C-V/m-sqN)

Temperature (C)

Figure 10. Variation of d31g31 of PZT-5A (Berlincourt et al., 2007)

0

1000

2000

3000

4000

5000

6000

0 100 200 300 400 500

Phasevelocity(m/s

)

S0 at 20C

A0 at 150C

A0 at 80C

Frequency (kHz)

A0 at 20C

S0 at 150C

S0 at 80C

Figure 11. Combined effect of changing aluminum elastic modulus(static) and thermal expansion on phase velocity.

58

60

62

64

66

68

70

0 20 40 60 80 100 120 140 160

Youngsmodulus(GPa)

Al staticAl dynamicPZT-5A

Temperature (C)

Figure 9. Variation of Youngs moduli (ANC-5 1955; Lord andOrkney, 2000; Williams et al., 2004).




9/17

three-dimensional GW field excited and sensed by

arbitrarily shaped, surface-bonded finite-dimensional

piezos in isotropic plates. The models assume uncoupled

dynamics and the piezo-actuator is modeled as causing

surface shear traction along its edge in the direction

normal to the edge on the surface of the underlying

substrate. More recently, extensions of these modelsto other structural configurations (isotropic beams,

hollow cylinders, and composite plates) have also been

proposed (Raghavan and Cesnik, 2007a,b). These

models have been validated extensively by numerical

and experimental results. The inputs for these models

are the material properties, dimensions of the structure,

the piezo dimensions, the piezo-actuators shear

traction magnitude, and piezo-sensors piezoelectric

constant and Youngs modulus, all of which are

thermally sensitive parameters as explained above.

Reduced transducer dimensions were used to generate

the theoretical plots. A 20% reduction in transducer size

over nominal values was assumed based on an earlierwork to correlate frequency response plots from

experiments with theoretical predictions for piezos of

the same thickness on similar aluminum plates

(Raghavan and Cesnik, 2005b). This is to account for

shear lag, which causes strain transfer to happen over a

small area close to the actuator edge, and not exactly

the actuator edge, as assumed in the pin-force model.

In addition, variations in the traction magnitude exerted

by the piezo due to changing elastic moduli and

dimensions were accounted for using the equation for

static actuation by piezos in Chaudhry and Rogers

(1994). These models were used to generate theoretical

predictions for the time-domain sensor response signalsat various temperatures between 20 and 1508C for the

configuration used in the experiments, incorporating

data for the thermal variation of the model inputs. With

that, it was examined whether the experimentally

observed increase in time-of-flight of the GW S0 mode

wave-packet received by Sensor 2 and the change in

sensor response peak-to-peak magnitude could be

captured. Spectrograms of these signals were generated

(for both the theoretical and experimental ones) and

the time-of-flight of the first transmitted S0 pulse was

computed from them. This also allowed examining to

see whether there was a change in the center frequency

of this pulse with temperature. The time-of-flight hereis defined as the time corresponding to the peak in the

first wave pulse (from the spectrogram) minus half

the excitation pulse span (corresponding to the peak of

the excitation), as shown in Figure 12. It was calculated

with a resolution of the experimental sampling rate, i.e.,

0.1ms. The comparison between theory and experiment

is shown in Figure 13 for time-of-flight and in Figure 14

for sensor response magnitude (peak-to-peak). To get

an estimate of the error in time-of-flight, the raw

un-averaged 30 signals were collected for two points

per thermal cycle (1008C while heating and 708C while

cooling). The data from Sensor 3 is very similar.

Discussion of Results and Implications for GW SHM

The theoretical estimates for time-of-flight are in

agreement with the experimental data (within error

(a)

(b)

(c)

10

0

10

0.05

200

150

100

0

0.05

Actuation

signal(V)

Sensor

signal(V)

Frequency

(kHz)

Time (s)

10

2

x104

x104

0.2 0.4

Time of flight

0.6 0.8

10.2 0.4 0.6 0.8

x10410.2 0.4 0.6 0.8

Figure 12. Extraction of time-of-flight at 908C while heating: (a)actuation signal; (b) Sensor 2 response signal; and (c) spectrogramof signal in (b).

0.02

0.04

0.06

0.08

0.1

0.12

0 20 40 60 80 100 120 140 160

Cycle 1 heating

Cycle 1 cooling

Cycle 2 heating

Cycle 2 cooling

Theoretical (static modulus variation)

Temperature (C)

Sensor2responsepeak-to-peakmagnitude

Figure 14. Variation in response magnitude (peak-to-peak) of firsttransmitted S0 mode received by Sensor 2.

19.5

20

20.5

21

21.5

22

0 20 40 60 80 100 120 140 160

Time-of-flight(s)

Exper imental (cycle 1 heat ing) Exper imental (cycle 1 cool ing)

Exper imental (cycle 2 heat ing) Exper imental (cycle 2 cool ing)

Theoretical (static elastic modulus) Theoretical (dynamic elastic modulus)

Temperature (C)

Figure 13. Variation in time-of-flight of first transmitted S0 modereceived by Sensor 2.




10/17

margins) and the agreement seems better for the

theoretical data set generated assuming static elastic

modulus variation. The center frequency was calculated

with a resolution of 3 kHz, which is limited by the time

span of the signal recorded. With this resolution, the

center frequency of the sensed S0 mode pulse remained

constant at 133 kHz for both the theoretical andexperimental data sets at different temperatures. For

the sensor response magnitude prediction, there is

clearly a significant gap between the theoretical estimate

and experimental data. The experimentally observed

increase of up to 33% while heating to 908C in the

response magnitude of Sensor 2 is not captured by the

theoretical model, but the decrease of 18% at 1508C is

predicted within the error margin. One possible expla-

nation for this is changing bond layer properties with

temperature (the model used here assumed perfect

bonding). Data were unavailable for variation of bond

layer elastic modulus with temperature and therefore it

could not be quantified. However, the static modulusvariation and thermal expansion data can be used to

give a good approximation to account for the slowing

down of wavespeeds with increasing temperature, and

is used in the subsequent section to generate damage

location estimates. An empirical compensation

approach is used for varying response magnitude.

From these experiments, it is clear that temperature

can cause significant changes in the magnitude and

time-of-flight for the first transmitted pulse received by

the sensor. Even with the instantaneous temperature

known, this causes greater error margins in the

magnitude and time-of-flight measurement under

slightly varying temperature. In most pitch-catchapproaches used, changes in these very features are

used to conclude whether damage is present in the

actuatorsensor path or not. It should be clarified that

in this context, pitch-catch approaches refer to those

that use a dense network of transducers and rely on

changes in the GW signal transmitted over the direct

actuatorsensor paths. Therefore, the pitch-catch

method is inherently more sensitive to false positives in

damage detection under varying temperature. On the

other hand, pulse-echo approaches typically rely on the

absence or presence of reflected or scattered echoes

between the actuation/first transmitted pulse and the

boundary reflection (although some researchers do useechoes arriving after the first boundary reflection).

If damage is present, regardless of temperature increase,

there will always be some GW reflection/scattering and

a sensor adjacent to the actuator or elsewhere should be

able to pick this up (assuming the damage is sensitive

to the mode and frequency of the incident GW).

Therefore, in principle, the only modification to make

for GW pulse-echo based damage characterization

under varying temperature would be to account for

varying GW speeds and scale the reflection magnitude

according to the changed sensor sensitivity at tha

particular temperature. However, complications arise

for pulse-echo methods in structures with features, such

as rivets in trying to detect and locate mild damage

which is roughly at the same distance (within a few cm

from the actuator as the rivet. In this context, damage is

called mild if the magnitude of the reflected GW by it iscomparable to that of the reflection from the rivet

structural discontinuity. This is explored experimentally

in the next section.

DAMAGE CHARACTERIZATION UNDER

VARYING TEMPERATURE

Baseline Signals and Threshold Values

As mentioned in the previous section, before any

damage was introduced two data sets were obtained for

the baseline, pristine condition of the specimen (whichhad transducers bonded with Epotek 353ND) from two

identical thermal cycles on different days. This was to

get a sense of the repeatability of the baseline condition

for a given temperature between the two cycles. The

baseline signal read by Sensor 1 (collocated with the

actuator) is shown in Figure 15. There is some non-zero

signal between the actuation pulse and the boundary

reflection due to mild A0 mode excitation and some

reflection from Sensors 2 and 3. At 208C, the A0 mode

reflection from Sensors 2 and 3 is discernible, while the

S0 mode reflection is negligible. However, at highe

temperatures, the S0 mode reflection from Sensors 2

and 3 becomes stronger (e.g., see Figure 15)Furthermore, there is some error in the magnitudes o

these mild reflections received at Sensor 1 when data

from the two cycles are compared. Figure 16 shows one

of the worse scenarios in this regard, while Figure 17 is

an example of better repeatability observed among

2 4 6 8 10

x 105

0.1

0.05

0

0.05

0.1

Time (s)

Sensor1respo

nse(V)

20C120C (heating)

S0mode actuation

Boundary reflections

Undesired A0

mode actuation

A0

reflectionfrom s2 and s3

S0

reflection

from s2 and s3

Figure 15. Signal read by Sensor 1 at 20 and 1108C (Cycle 1) fopristine condition.




11/17

the readings. This error in repeatability arises from the

variation in time-of-flight and sensor response magni-tude under changing temperature. As mentioned earlier,

there is a 18C error in the temperature read by the

thermocouples and the autoclave continues heating/

cooling at 18C/min during the 30-s period of data

collection for each temperature reading. With large

temperature-change rates, this problem is further

exacerbated. This explains why poor signal repeatability

was observed in the initial experiments in the oven. The

variation between the two baseline cycles defines the

threshold values for the subsequent damage detection

experiments. In practice, it would be advisable to get

some more data for the variability in baseline condition

at various temperatures, particularly in less homoge-neous structural layouts with rivets, stiffeners, etc. In

addition, while data were collected here at 108C

intervals, it can also be used for baseline interpolation

at intermediate temperatures. This can be done by

simple weighted averaging of the two signals taken at

the multiples of 108C within which the intermediate

temperature lies (see Figure 18). Of course, this is

possible because in this case, the variation between

signals at 108C intervals is not significant enough

to cause destructive interference between them

when averaged. However, if the piezo is used to monitorlarger structures or using higher frequencies, the spacing

between reflections from structural features may

increase in the time domain and destructive interference

might become a possibility with weighted averaging. In

such cases, readings can be taken at more closely spaced

temperature intervals to overcome this issue.

Once the two sets of baseline signals were recorded,

threshold energy levels were obtained from them.

The threshold energy in this context is defined as the

peak signal energy obtained from the spectrogram of the

difference signal between the two baselines (scaled for

18V actuation as well as to match the peak-to-peak

magnitude of the first S0 actuation pulse between thetwo) at the same temperature. The threshold energy is

chosen to be the peak value from the spectrogram in the

time window between the end of the S0 mode actuation

signal seen at Sensor 1 and the beginning of the

boundary reflections (illustrated in Figure 16 and

Figure 17). As explained in the authors previous work

(Raghavan and Cesnik, 2007d), some kind of time

frequency analysis is essential for GW signals because

they allow for distinguishing among the different GW

modes possible and also enable tracking the change in

the center frequency of the scattered GW pulse (since

damage may be more sensitive to frequencies other than

the center frequency of the incident GW pulse). This cansignificantly affect the speed of the scattered pulse,

which can contribute to location estimate errors.

Indentation Damage

Subsequently, damage was introduced artificially in

the plate by drilling. As alluded to at the end of an

earlier section, the case of mild damage was first

explored. To ensure that the reflections were not too

strong, damage was introduced along the axis of weaker

0 2 4 6 8 10

105

0.15

0.1

0.05

0

0.05

0.1

0.15

Sensor1r

esponse(V)

90C85C (Interpolated)80C

Time (s)

Figure 18. Illustration of interpolated signal (at 858C) obtained byweighted averaging of recorded signals at 80 and 908C (slightlyoffset by a constant value for clarity) from pristine specimen whileheating during Cycle 1.

Sensor1

response(V)

Frequency

(kHz)

0.05

160

140

12010080

60

2 4 6

Time (s)

Cycle 1 (pristine)Cycle 2 (pristine)Difference

8

8 x 103

6

4

2

10

2 4 6 8 10

x 105

x 105

180

0.05

0

Thresholdenergyvalue

Time-frequencywindow forthresholdenergy

Figure 17. Sensor 1 response during Cycles 1 and 2 for pristinecondition along with the difference signal and its spectrogram at 608Cwhile cooling (signals offset by a small DC voltage for clarity).

Sensor1

response(V)

Frequency

(kHz)

0.05

0.05

0

150

50

100

2 4 6 8 10

0.15

Cycle 1 (pristine)Cycle 2 (pristine)Difference

0.05

0

0.1

x 105

2 4 6

Time (s)

8 10

x 105

Time-

frequencywindow forthresholdenergy

Thresholdenergyvalue

Figure 16. Sensor 1 response during Cycles 1 and 2 for pristinecondition along with the difference signal and its spectrogram at1208C while heating (signals offset by a small DC voltage for clarity).




12/17

actuation of the rectangular actuators. Theoretical

models developed earlier by the authors (Raghavan

and Cesnik, 2005b) predicted a 30% weaker GW strain

field along this direction. Experiments were first donefor a triangular cross-sectional indentation of maximum

diameter 5 mm and depth 1.7 mm (Figure 19(a)). The

damage center is 8 cm away from the plate center and its

location is shown in Figure 4. The signal obtained from

Sensor 1 at room temperature after introducing this

damage is shown in Figure 20, along with the pristine

condition signal and the difference between the two.

As a well-established practice in the GW SHM commu-

nity, this difference signal is used for damage detection

and characterization. There are clear reflections in the

difference signal whose peak energies from the spectro-

gram are well above the energy threshold defined for

208C. The two distinct reflections in the signal are theS0- and A0-mode reflections from the indentation.

As explained in the authors earlier work (Raghavan

and Cesnik, 2007d), even though the S0 mode is

predominantly excited, mode conversion is possible at

a defect. In this case, because of the controlled

introduction of damage in the specimen, the A0 and S0mode reflections were expected and their time-of-flight

values could be estimated. However, in field applica-

tions, one has to blindly infer all information from the

GW signals. Therefore, for practical situations, one has

to be able to identify the GW mode of each reflection as

well. This can be done from the reflections character-

istics in the timefrequency plane. For the A0 mode,since higher frequencies travel much faster than the

lower ones, the peaks at higher frequencies within the A0mode reflection arrive faster than those for the lower

ones in the timefrequency plane. On the other hand,

for the S0 mode the converse is true, but the variation in

S0 mode wavespeeds is milder over the frequency range

and structure used in these experiments. Once the mode

is identified, the group speed(s) corresponding to the

frequency at which the peak within the reflection occurs

is used with the time-of-flight for estimating the radial

location of the damage site (see Raghavan and Cesnik

2007d). As mentioned in an earlier section, the group

speeds for a particular temperature are computed by

accounting for the Youngs modulus at that temperatureand thermal expansion in the dispersion curves. In this

case, the GW packet originates from one edge of the

actuator and its echo travels back all the way to the

other edge and has to further traverse a distance equal to

half the sensor dimension before the peak of the

reflected GW is seen in the sensor response. Therefore

minor correction terms to account for the transi

times of the GW packet through the transducers are

subtracted from these estimates. At 208C, the radia

location estimates of the indentation are 7.2 and 8.2 cm

(from the plate center) based on the S0 and A0 mode

reflections, respectively. The analysis done in thi

study was done by manually tracking the reflections inthe spectrograms. In principle, the chirplet matching

pursuit algorithm used earlier by the authors (Raghavan

and Cesnik, 2007d) should allow automated tracking

and better resolution for real-time processing in

end applications. However, its implementation in

LastWave 2.0 (Bacry, 2007) makes approximations to

reduce computational complexity which do no

capture the slowing down of wavespeeds at elevated

temperatures seen in Figure 13.

The indented specimen was thermally cycled in the

autoclave to check whether the difference signa

remained above the pre-defined threshold level at each

temperature point. Some of the signals read by Sensor 1during this experiment are shown in Figure 20 and

Figure 21. The results are also summarized in Table 1

The A0 and S0 mode reflections from the indentation

had peak energy (from the spectrogram, in the excited

frequency band as shown in Figure 20 for the signal at

208C) well above the threshold up to 808C while heating

Some of the S0 mode reflections (from 50 to 808C) mixed

with the excitation difference signal, due to which the

S0 mode reflection underestimated the damage location

(shown in Figure 21). Beyond 808C, the A0 mode

(a) (b)

Figure 19. Photographs of damage introduced: (a) indentation and(b) through-hole.

Sensor1

response(V)

Frequency

(kHz)

0.05

0

2 4 6 8 10

2 4 6

Time (s)

PristineIndentation

Difference

8 10

0.2

0.1

0

x 105

x 105

0.05

180

160140

120

100

80

60

S0

modereflection

A0

modereflection

Figure 20. Sensor 1 response at 208C for pristine and indented specimens, along with the difference signal (offset by smaDC values for clarity) and its spectrogram.




13/17

Table

1.

Summaryo

fresu

lts

show

ing

trends

in

thermalexperimentfordamage

cha

racterization

with

indented

specimen

(Key:

Temp.

Temperature;

(h)

heating

phase;

(c)

cooling

phase;

S0

ToF

time-o

f-flightfor

S0

mode

reflection

from

indentation;

S0

loc.

est.

radiallocatione

stimate(relativetoplatecenter)ofdamagebasedonS

0

modereflectio

n).

Temp.

(8C)

S0

ToF(ks)

S0

energy(V2)

Threshold

energy(V2)

S0

above

threshold?

S0

loc.est.

(cm)

A0

ToF(ks)

A0

energy(V2)

A0

above

threshold?

A0loc.est.(cm)

20

(h)

28.1

0

0.0

631

0.0

142

Yes

7.2

50.6

5

0.2

8825

Yes

8.2

40

(h)

28.0

0

0.1

229

0.0

025

Yes

7.2

51.2

0

0.3

695

Yes

8.6

70

(h)

18.6

0

0.0

739

0.0

154

Yes

4.7

54.6

0

0.5

895

Yes

9.1

110

(h)

31.2

0

0.1

808

0.1

850

No

7.9

56.0

0

0.1

959

Yes

9.2

130

(h)

18.6

0

0.0

66

0.0

517

Yes

4.5

34.1

0

0.0

886

Yes

4.9

150

0.0

275

25.5

0

0.0

839

Yes

4.0

120

(c)

0.0

375

28.6

0

0.1

363

Yes

4.1

80

(c)

25.5

0

0.2

688

0.0

525

Yes

6.5

56.8

0

0.1

502

Yes

9.5

50

(c)

27.5

0

0.0

681

0.0

033

Yes

7.1

55.8

2

0.2

240

Yes

9.4

20

(c)

30.3

5

0.0

483

0.0

023

Yes

7.8

53.9

5

0.4

228

Yes

8.8




14/17

reflection was still above the threshold, but had peakenergy that was comparable to the threshold. The

weaker S0 mode reflection had peak energy lower than

the threshold at some points after 908C while heating.

In addition, as illustrated in Figure 21, at some points

over 1008C, there is a reflection that arrives approxi-

mately where the S0 mode was seen at lower tempera-

tures, but can be wrongly identified as A0 mode by its

timefrequency characteristics. This results in much

larger errors for the radial location estimate.

Subsequently, while cooling below 808C, again both

reflections were well above the threshold, and gave

reasonable location estimates.

Through-hole Damage

Following this, a through hole of diameter 7.5 mm

was drilled at the same location (Figure 19(b)). The

response of Sensor 1 recorded after drilling through at

208C is shown in Figure 22. As seen in the difference

signal, the S0 mode reflection is now much stronger

while the A0 mode reflection appears much weaker

(but still well above the threshold at 208C). The radial

location estimates are 8.6 and 8.3 cm based on the S0 and

A0 mode reflections, respectively at 208C. The results

from thermally cycling the specimen with the through-

hole are shown in Figures 22 and 23 and tabulated inTable 2. In this case, the S0 mode reflection is reasonably

above the threshold at all temperatures. There is an

error of 1.9 cm in the estimate (based on the S0 mode) at

1508C, where the signal is the weakest. At a couple of

points, the radial location of the damage is over-

estimated by more than 1 cm (1008C while heating and

508C while cooling), which is possibly due to the mixing

of the S0 mode with the difference in the reflection from

Sensors 2 and 3. At other temperatures, the radial

location estimates based on the S0 mode are within 1 cm.

The A0 mode reflection, which was weak at 208C to

begin with, is discernible up to 708C while heating, bu

beyond that is indistinguishable from the difference in

boundary reflection until the specimen cools back to

room temperature.

DISCUSSION

Thus, for mild damage up to 808C, detection was no

problematic, but there was a slightly increased error in

location as temperature increased. However, beyond

that temperature, there is a definite decrease in

sensitivity, as reflected in the poorer detection

characterization capability in the indentation experi

ment. This can be attributed to the higher sensitivity

of the substrate elastic modulus to temperature

at higher temperatures (Figure 9), causing greater

variation in the mild reflections from Sensors 2 and 3

2 4 6 8 10

(a)

Pristine

Indentation

Difference

1 2 3 4 5 6 7 8 9 10 11

x 105

x 10

5

0.1

0.05

0

0.05

0.1

0.1

0.05

0

0.05

0.1

Pristine

Indentation

Difference

Sensor1

response(V)

S0

reflection mixed with excitation

A0

reflection

Sensor1

response(V)

A0

reflectionnegligible

Time (s)

(b)S0

reflection appearslike A

0mode in t-f ane

Figure 21. Sensor 1 response for pristine and indented specimens,along with the difference signal (offset by small DC values for clarity)at (a) 608C during heating and (b) 1508C.

0.1

0.05

0

0.05

2 4 6 8 10

0

(b)

PristineHoleDifference

PristineHoleDifference

0.05

0.05

0.1

x 105

2 4 6 8 10

(a) x 105

Figure 23. Sensor 1 response during heating for pristine andthrough-hole specimens, along with the difference signal (offset bysmall DC values for clarity) at: (a) 70 and (b) 1408C.

Sensor1

response(V)

Frequency

(kHz)

0.05

0.05

0

S0

modereflection

A0

modereflection

Pristine

Hole

Difference

Time (s)

280

100

120

140

160

4 6 8 100

0.05

0.1

0.15

0.2

2 4 6 8 10

x 105

x 105

Figure 22. Sensor 1 response at 208C for pristine and through-hole specimens, along with the difference signal (offset by smaDC values for clarity) and its spectrogram.




15/17

Table2.

Summaryofresu

ltsshow

ingtrendsinthermalexperiment

fordamagecharacterizationusings

pecimenw

iththrough-h

ole(Key:

Te

mp.

Temperature;

(h)

heatingphase;

(c)

coolingphase;

S0

ToF

time-o

f-flightfo

rS

0

modereflectionfrom

thru-h

ole;

S0

loc.

est.

radiallocationestima

te(relativetoplate

center)ofdamagebasedonS

0

modereflection;

Bndry

A0

mode

reflectionpeakw

ithinboundaryreflection).

Temp.

(8C)

S0

ToF(ks)

S

0

energy(V2)

Threshold

energy(V2)

S0

above

threshold?

S0

loc.est.(cm)

A0

ToF

(ks)

A0

energy(V2)

A0

above

threshold?

A0

loc.est.(cm)

20

(h)

33.4

5

0.2

198

0.0

142

Y

es

8.6

49.7

5

0.0

763

Yes

8.3

40

(h)

34.3

0

0.2

972

0.0

025

Y

es

8.8

55.2

0

0.1

119

Yes

9.0

70

(h)

32.5

0

0.2

463

0.0

154

Y

es

8.3

56.7

0

0.2

128

Yes

9.0

110

(h)

40.6

0

0.2

578

0.1

850

Y

es

10.2

Bnd

ry

130

(h)

33.4

0

0.3

105

0.0

517

Y

es

8.2

Bnd

ry

150

25.6

0

0.1

073

0.0

275

Y

es

6.1

Bnd

ry

120

(c)

31.0

0

0.1

997

0.0

375

Y

es

7.6

Bnd

ry

80

(c)

28.0

0

0.2

985

0.0

525

Y

es

7.0

Bnd

ry

50

(c)

36.2

0

0.3

393

0.0

033

Y

es

9.3

Bnd

ry

20

(c)

33.8

0

0.2

018

0.0

023

Y

es

8.7

53.6

0

0.0

623

Yes

8.8




16/17

and consequently, poorer repeatability of the

baseline signals. In addition, the sensor sensitivity

(in terms of signal magnitude) drops below the value

at 208C beyond 1308C (as seen in Figure 14). For the

through-hole, which can be termed moderate damage,

detection was clearly possible at all temperatures, but

at a few points (3 of a total of 29 cases), there wasinaccuracy in the location estimate (by up to 2.2 cm for

damage located at 8 cm), partly due to interference with

the difference in reflection from Sensors 2 and 3

(simulating structural features in field applications).

One way to reduce the error in location is to use higher

center frequencies and/or fewer number of cycles (which

increases frequency bandwidth) in the actuation signal.

However, that was not feasible in the present setup

(which did not use amplifiers), since significant distor-

tion was observed for such actuation signals at higher

temperatures due to increasing actuator capacitance.

In addition, in the present experimental setup, the

scaling of the signals to compensate for changingactuation level increased the signal-to-noise ratio at

higher temperatures. This indicates the need for reliable

actuation signal amplification for actuating piezos at

elevated temperatures.

SUMMARY AND CONCLUSIONS

This study addressed the issue of GW SHM using

piezos under thermal variations as expected in spacecraft

internal structures. Experiments were done to determine

a bonding agent (for piezos on aluminum plates) that

did not degrade under thermal variations from 20 to1508C. Using this bonding agent (Epotek 353ND),

results from controlled experiments done to examine

changes in GW propagation and transduction using

PZT-5A piezos under quasi-statically varying tempera-

ture (also from 20 to 1508C) were presented. Thermally

sensitive variables in the experiments were identified and

quantified to model the experimentally observed

changes under temperature variation. The Youngs

modulus of the structure was the most critical of these

parameters. The increase in time-of-flight of GW pulses

with increasing temperature was captured by the model

(within the error margins). However, there was a

significant gap in the prediction of the large increase insensor response magnitude up to 1008C. The stronger

vulnerability of pitch-catch approaches (that detect

damage in the direct actuatorsensor path based on

changes in the first pulse transmitted over this path) to

false positives under changing temperature was then

explained. Finally, detection and location of damage

(by drilling) using the pulse-echo GW testing approach

in the presence of mild structural GW scatterers

(to simulate rivets) was explored in the same tempera-

ture range. Damage characterization of a half-plate

thickness indentation at 8 cm from the actuators was no

significantly affected up to 808C, but beyond tha

temperature, detection/characterization was difficult

The problems beyond 808C can be traced to increased

sensitivity of substrate elastic modulus to temperature

and weaker sensor sensitivity beyond 1308C. For a

through-hole, damage detection and characterizationwas possible at all temperatures and, except at a few

temperatures (3 out of 29), damage was located within

1 cm for a nominal location of 8 cm and, hole diamete

0.75cm. Suggested approaches for improving the

sensitivity at higher temperatures include testing a

higher frequencies and/or with shorter time-span

excitation pulses, with reliable, low-distortion actuation

amplifiers.

ACKNOWLEDGMENTS

The authors gratefully acknowledge Dr. W.H. Prosser

(NASA Langley) for suggesting the investigation o

b o nd in g agen t d egr ad ati on . T he ass i st ance o

Mr. Danny Lau (University of Michigan) in setting up

one of the experiments is also appreciated. This work

was supported by the Space Vehicle Technology

Institute under grant NCC3-989 jointly funded by

NASA and DoD within the NASA Constellation

University Institutes Project, with Ms. Claudia Meye

as the project manager.

REFERENCES

Bacry, E. 2007. LastWave 2.0 software, http://www.cmap.polytechniquefr/$ bacry/LastWave

Berlincourt, D., Krueger, H.H.A. and Near, C. 2007. Properties oMorgan Electroceramics, Morgan Electroceramics TechnicaPublication TP-226, http://www.morganelectroceramics.com

Blaise, E. and Chang, F.-K. 2001. Built-in Diagnostics foDebonding in Sandwich Structures under ExtremeTemperatures, Proceedings of the 3rd International Workshopon Structural Health Monitoring, pp. 154163, StanfordCalifornia, USA.

Carlos, G.-C., Alan, C. and Gordon, A. 2005. Health Managemenand Automation for Future Space Systems, Proceedings of the

AIAA Space 2005 Conference and Exposition, pp. 15411557Long Beach, CA.

Chambers, J., Wardle, B. and Kessler, S. 2006. DurabilityAssessment of Lamb-wave Based Structural Health MonitoringNodes, Proc. 47th AIAA/ASME/ASCE/AHS/ASC StructuresStructural Dynamics, and Materials Conference, Paper # AIAA2006-2263, Newport, Rhode Island.

Chaudhry, Z. and Rogers, C.A. 1994. The Pin-force ModeRevisited, Journal of Intelligent Material Systems andStructures, 5(3):347354.

Derriso, M.M., Olson, S., Braisted, W., DeSimio, M., Rosenstengel, Jand Brown, K. 2004. Detection of Fastener Failure in aThermal Protection System, Proceedings of the SPIE, 5390585596.




17/17

Ecke, W., Latka, I., Willsch, R., Reutlinger, A. and Graue, R. 2001.Fiber Optic Sensor Network for Spacecraft HealthMonitoring, Measurement Science and Technology, 12:974980.

Epoxies Etc. 2006. 10-3004 Adhesive Technical Datasheet, http://www.epoxies.com, Cranston, RI.

Epoxy Technology 2006. Epoxies 301 and 353ND TechnicalDatasheets, http://www.epotek.com, Billerica, MA.

Friebele, E.J., Askins, C.G., Bosse, A.B., Kersey, A.D., Patrick, H.J.,

Pogue, W.R., Putnam, M.A., Simon, W.R., Tasker, F.A.,Vincent, W.S. and Vohra, S.T. 1999. Optical Fiber Sensorsfor Spacecraft Applications, Smart Materials and Structures,8:813838.

Hilton, H.H. and Vail, C.F. 1993. Bending-torsion Flutter of LinearViscoelastic Wings Including Structural Damping, Proc.AIAA/ASME Structures, Structural Dynamics and MaterialsConference, Paper # AIAA-93-1475-CP, La Jolla, CA.

Huang, J., Rose, J., Gordon, J. and Boucher, R. 2005. StructuralSensor Testing for Space Applications, Proceedings of the SPIE,5762:8899.

Konstantinidis, G., Drinkwater, B.W. and Wilcox, P.D. 2006. TheTemperature Stability of Guided Wave Structural HealthMonitoring Systems, Smart Materials and Structures,15:967976.

Larson, W.J. and Wertz, J.R. 1995. Space System Design and Analysis,

2nd edn, Microcosm, CA/Kluwer Academic, The Netherlands.Lee, B.C., Manson, B. and Staszewski, W.J. 2003. Environmental

Effects on Lamb Wave Responses from Piezoceramic Sensors,Materials Science Forum, 440441:195202.

Lin, M., Kumar, A., Qing, X., Beard, B.J., Russell, S.S., Walker, J.L.and Delay, T.K. 2003. Monitoring the Integrity of FilamentWound Structures by Built-in Sensor Networks, Proceedings ofthe SPIE, 5054:222229.

Lord, J.D. and Orkney, L.P. 2000. Elevated Temperature ModulusMeasurements using the Impulse Excitation Technique,National Physical Laboratory Measurement Note,Teddington, UK.

Lu, Y. and Michaels, J.E. 2005. A Methodology for StructuralHealth Monitoring with Diffuse Ultrasonic Waves in thePresence of Temperature Variations, Ultrasonics, 43:713731.

Matweb 2007. Material property database, http://www.matweb.com

Myers, D.E., Martin, C.J. and Blosser, M.L. 2000. ParametricWeight Comparison of Advanced Metallic, Ceramic Tile, andCeramic Blanket Thermal Protection Systems, NASA-TM-2000-210289, NASA Center for AeroSpace Information.

Raghavan, A. and Cesnik, C.E.S. 2005a. Lamb Wave Methods inStructural Health Monitoring, In: Inman, D., Farrar, C.R.,Lopes, Jr, V. and Steffen, Jr, V. (eds), Damage Prognosis, JohnWiley & Sons, UK.

Raghavan, A. and Cesnik, C.E.S. 2005b. Finite DimensionalPiezoelectric Transducer Modeling for Guided Wave BasedStructural Health Monitoring, Smart Materials and Structures,

14:14481461.

Raghavan, A., and Cesnik, C.E.S. 2007a. 3-D Elasticity-basedModeling of Anisotropic Piezocomposite Transducers forGuided-wave Structural Health Monitoring, Journal of

Vibration and Acoustics (Transactions of the ASME), SpecialIssue on Damage Detection and Structural Health Monitoring129(6):739751.

Raghavan, A. and Cesnik, C.E.S. 2007b. Modeling of Guided-waveExcitation by Finite-Dimensional Piezoelectric Transducers inComposite Plates, Presented at the 15th AIAA/ASME/ASCAdaptive Structures Conference, Paper 2007-1725, Honolulu,Hawaii, Apr. 2326.

Raghavan, A. and Cesnik, C.E.S. 2007c. Review of Guided-waveStructural Health Monitoring, The Shock and Vibration Digest,39:91114.

Raghavan, A. and Cesnik, C.E.S. 2007d. Guided-wave SignalProcessing using Chirplet Matching Pursuits and Mode

Correlation for Structural Health Monitoring, SmartMaterials and Structures, 16:355366.

Schulz, M.J., Sundaresan, M.J., Mcmichael, J., Clayton, D., Sadler, R.and Nagel, B. 2003. Piezoelectric Materials at ElevatedTemperature, Journal of Intelligent Material Systems andStructures, 14(11):693705.

Stanley, D., Cook, S. and Connolly, J. (study managers) 2005.Exploration Systems Architecture Study Final Report, NASA-TM-2005-214062, NASA Space Transportation InformationNetwork.

U.S. Munitions Board Aircraft Committee 1955. Strength ofMetal Aircraft Elements, ANC-5 Bulletin, Republican Press,Hamilton, Ohio.

Williams, B.R., Inman, D.J. and Wilkie, W.K. 2004.Temperature-dependent Thermoelastic Properties for MacroFiber Composite Actuators, Journal of Thermal Stresses,

27:903915.Yang, J., Chang, F.-K. and Derriso, M.M. 2003. Design of a Built-in

Health Monitoring System for Thermal Protection Panels,Proceedings of the SPIE, 5046:5970.


Effects of Elevated Temperature on Guided-Wave

Documents

Transcript of Effects of Elevated Temperature on Guided-Wave