aerospace

Article
S-Parameter-Based Defect Localization for Ultrasonic
Guided Wave SHM †
Gift Nyikayaramba * and Boris Murmann
Murmann Mixed-Signal Group, Department of Electrical Engineering, Stanford University,
330 Jane Stanford Way, Stanford, CA 94305, USA; murmann@stanford.edu
* Correspondence: giftn@stanford.edu; Tel.: +1-919-282-7211
† This paper is an extended version of our conference paper published at the 12th International Workshop on
Structural Health Monitoring (IWSHM), Stanford, California, USA, 10–12 September 2019, which was
selected for and invited to this special issue.




Received: 28 February 2020; Accepted: 19 March 2020; Published: 20 March 2020


Abstract: In this work, an approach for enabling miniaturized, low-voltage hardware for active
structural health monitoring (SHM) based on ultrasonic guided waves is investigated. The proposed
technique relies on S-parameter measurements instead of time-domain pulsing and thereby trades
off longer measurement times with lower actuation voltages for improved compatibility with dense
complementary metal-oxide-semiconductor (CMOS) chip integration. To demonstrate the feasibility
of this method, we present results showing the successful localization of defects in aluminum
and carbon-fiber-reinforced polymer (CFRP) test structures using S-parameter measurements.
The S-parameter measurements were made on benchtop vector network analyzers that actuate
the piezoelectric transducers at output voltage amplitudes as low as 1.264 Vpp .

Keywords: structural health monitoring; ultrasonic guided waves; electronics; integrated circuits;
signal processing; S-parameters; signal-to-noise ratio

1. Introduction
The increased use of high-performance composites and the need to lower maintenance costs while
ensuring high standards of safety and reliability are prevailing trends in the aerospace industry. Together,
these trends are fueling a shift from a schedule-based maintenance approach to a condition-based
one [1,2]. In-situ structural health monitoring (SHM) systems are a key enabler of this condition-based
maintenance paradigm because they can test the physical condition of structures in an automated,
on-demand, and real-time fashion. Ultrasonic guided waves (UGWs) are a widely used modality in the
implementation of SHM systems because they can travel long distances in thin-walled structures with
minimal amplitude loss [1]. Furthermore, the transducers that excite UGWs can be easily incorporated
into new and existing structures [3–5].
A key bottleneck in the deployment of in-situ UGW SHM systems is the use of data acquisition
electronic systems that are bulky and unsuited for wide scale operation in the aircraft environment,
which is highly sensitive to additional weight and volume. The electronic systems are large because
they actuate the piezoelectric transducers that emit UGWs using sequences of short tone-burst pulses
so as to minimize dispersion and maximize resolution [1]. To attain a receiver signal-to-noise ratio
(SNR) that is good enough for the collected data to provide accurate defect detection and localization
information, these tone-burst signals must have large amplitudes (~several tens to hundreds of volts)
so as to have adequate energy to be distinguishable from noise. As a result, to withstand the large
electric fields that are generated by these pulses, ultrasonic SHM systems are typically implemented
using standard laboratory equipment [6–8] or bulky discrete electronic components [9,10].

Aerospace 2020, 7, 33; doi:10.3390/aerospace7030033 www.mdpi.com/journal/aerospace

Aerospace 2020, 7, 33 2 of 15
Aerospace 2020, 7, x FOR PEER REVIEW 2 of 15

An
Aneffective
effective avenue
avenue towards
towardsthe miniaturization
the miniaturization of SHM of electronics is the useisofthe
SHM electronics custom
use integrated
of custom
circuits
integrated (ICs). In prior
circuits work
(ICs). In in this work
prior area, our teamarea,
in this presented
our team a highly linear and
presented energy-efficient
a highly linear and
four-channel
energy-efficient piezo driver integrated
four-channel piezocircuit
driverfor UGW SHM
integrated [11].for
circuit However,
UGW SHM this chip
[11].requires
However, an this
external
chip
±36 V power
requires supply.±36
an external ThisV limits
powerthe chip’sThis
supply. scalability
limits the and necessitates
chip’s scalabilitytheandusenecessitates
of an expensive and
the use of
area-inefficient 0.25 µm Bipolar-CMOS-DMOS fabrication process.
an expensive and area-inefficient 0.25 μm Bipolar-CMOS-DMOS fabrication process. In more recent In more recent work, Tang et al.
demonstrated
work, Tang eta compact single-channel
al. demonstrated 12.7 Vsingle-channel
a compact pp transceiver for 12.7SHMVpp [12,13].
transceiver However,
for SHM this design
[12,13].
still requires
However, an design
this external high-voltage
still requires anpowerexternal amplifier for typical
high-voltage power useamplifier
cases. for typical use cases.
This
This work
work is is inspired by the 2014 2014 work
work of of Huang
Huangand andBednorz,
Bednorz,which whichintroduced
introducedthe the
useuse of
of scattering
scattering parameters
parameters (S-parameters)
(S-parameters) for UGWfor UGWSHM [14].SHMUsing [14]. measurements
Using measurements made on amade on
benchtop
avector
benchtop vector analyzer
network network analyzer
(VNA), they(VNA), they showed
showed that ultrasound
that ultrasound pitch-catch,
pitch-catch, pulse-echo,
pulse-echo, and
and electro-mechanical
electro-mechanical impedance
impedance signals
signals can be can be recovered
recovered from from the S-parameters
the S-parameters of a structure
of a structure under
under test. signals
test. These These have
signals have SNR
a higher a higher SNR and
and larger largerrange
dynamic dynamic range
than the than the corresponding
corresponding time-domain
time-domain
signals measured signals onmeasured on an oscilloscope,
an oscilloscope, and thus require and thus require significantly
significantly smaller actuation smallervoltages.
actuation In
voltages.
follow-on In follow-on
work, it waswork,also it was also established
established that S-parameter that S-parameter
measurements measurements allow for
allow for convenient
convenient
time-frequency time-frequency
analysis foranalysis for SHM [15,16].
SHM [15,16].
Our
Our recent
recent conference
conference publication
publication demonstrated,
demonstrated, for for the
the first
first time,
time, the the use
use of of S-parameter
S-parameter
measurements
measurementsfor forthe localization
the localization of defects
of defectsin aninaluminum
an aluminum test structure [17]. In[17].
test structure this In
paper,
this we extend
paper, we
extend
this work this
to work
also showto also show S-parameter-based
S-parameter-based defect localization
defect localization in a carbon-fiber-reinforced
in a carbon-fiber-reinforced polymer
(CFRP)
polymer test(CFRP)
structuretestand structure
then leverage andthe insights
then we gain
leverage theto propose
insights a we frequency-domain-based
gain to propose a
frequency-domain-based
on-chip system for ultrasonic on-chip
SHM.system
The rest forofultrasonic
this paperSHM. The restasoffollows.
is organized this paper is organized
Section 2 provides as
follows.
some Section on
background 2 S-parameters
provides some andbackground on S-parameters
their use in ultrasonic guided-wave and SHM. their Focus
use inis placed
ultrasonicon
guided-wave
explaining how SHM.
VNAs Focus
can isexploit
placedlonger
on explaining
measurement how VNAstimescan exploit
to make highlongerSNRmeasurement
measurements timesat
to make
low output high SNR
signal measurements
amplitudes at low
that are output with
compatible signalCMOS
amplitudes that aretechnology.
IC fabrication compatible In with CMOS
Section 3,
IC experimental
the fabrication technology.
setup that was In Section 3, the experimental
used to validate setup thatmeasurements
the use of S-parameter was used to validate the use of
in the localization
S-parameter
of measurements
defects in aluminum in thetest
and CFRP localization
structures ofisdefects
described.in aluminum
The results and of CFRP test structures
these experiments areis
described.in
presented The results
Section 4.ofFinally,
these experiments are presented
Section 5 discusses in Section
the future 4. Finally,
direction of thisSection
work and 5 discusses
concludes the
future
the direction of this work and concludes the paper.
paper.

2.2. Background
Background
S-parameters
S-parametersare widely
are widely used in electrical
used engineering
in electrical to describe
engineering the behavior
to describe the of linear, multi-port
behavior of linear,
electrical
multi-portnetworks
electricalinnetworks
response in to stimulation
response to by electrical signals.
stimulation Figure
by electrical 1 shows
signals. the S-parameter
Figure 1 shows the
representation of a two-port electrical
S-parameter representation network.electrical
of a two-port Four S-parameters
network. represent the reflection
Four S-parameters coefficients
represent the
at Ports 1 and
reflection 2 (S11 and
coefficients at S 22 ) and
Ports the2transmission
1 and coefficients
(S11 and S22) and (gains) from
the transmission Port 1 to(gains)
coefficients Port 2 from
(S21 ) and
Port
Port 2 to Port
1 to Port 2 (S211) and
(S12 ).Port
In this diagram,
2 to Port 1 (S12).a1Inand
thisa2diagram,
are the input
a1 andsignals at Ports
a2 are the input 1signals
and 2,atrespectively.
Ports 1 and
Similarly, b and
2, respectively.
1 b are the
Similarly,
2 reflected
b1 and b2 are the reflected signals at those ports. Since of
signals at those ports. Since the inputs and outputs aninputs
the ultrasonic
and
SHM
outputssystem
of anare electrical
ultrasonic SHM signals,
system thearepaths between
electrical transducers
signals, the pathsthrough
betweenthe structure through
transducers can be
characterized
the structure canusing beS-parameters
characterizedifusing the Lamb waves travelling
S-parameters if the Lamb on waves
them exhibit properties
travelling on themthat are
exhibit
linear and time-invariant.
properties that are linear and time-invariant.
Port 1 Port 2

a1 b2 1 2
S21 11 = 21 =
1 2 =0
1 2 =0
S11 S22
2 1
S12 22 = 12 =
2 1 =0
2 1 =0
b1 a2

Z0 Z0

Figure 1. S-parameter representation of a two-port network.

Figure 1. S-parameter representation of a two-port network.

From signal processing, we know that a signal can be represented both in the time-domain and
the frequency-domain. Even though they emphasize different properties of the signal, these
Aerospace 2020, 7, 33 3 of 15

Aerospace
From2020, 7, x FOR
signal PEER REVIEW
processing, 3 ofthe
we know that a signal can be represented both in the time-domain and 15
frequency-domain. Even though they emphasize different properties of the signal, these representations
representations
are equivalent, and are each
equivalent,
containsand each
all the contains all
information thatthe
weinformation
need to fullythat we needthe
characterize to signal.
fully
characterize the signal. Furthermore, a signal can be translated between
Furthermore, a signal can be translated between these two equivalent representations without these two equivalent
representations
information losswithout informationapplying
by appropriately loss by appropriately applying the
the Fourier Transform andFourier Transform
its inverse to theand its
signal.
inverse
Figure 2toshows
the signal. Figure 2 shows
a comparison between a comparison
the traditionalbetween the traditional
time-domain approach time-domain
to ultrasonic approach
SHM and to
ultrasonic SHM and the S-parameter approach. As shown, Fourier Transform
the S-parameter approach. As shown, Fourier Transform methods enable the use of S-parameters methods enable the
use of S-parameters
for SHM while allowing for SHM while
for the allowing for
maintained usethe
of maintained
well-established use of[18–20]
well-established
and emerging [18–20] and
[21–23]
emerging [21–23] time-domain-based algorithms for
time-domain-based algorithms for detecting and localizing defects. detecting and localizing defects.
The
The voltage
voltage amplitudes
amplitudes applied
applied to to the
the lead
lead zirconate
zirconate titanate
titanate (PZT)
(PZT) piezoelectric
piezoelectric transducers
transducers in in
the measurement of S-parameters for SHM in prior work [14] are of particular
the measurement of S-parameters for SHM in prior work [14] are of particular interest to our efforts to interest to our efforts
to implement
implement customcustom integrated
integrated circuitscircuits for ultrasonic
for ultrasonic SHM. In that SHM.work, In S-parameter
that work,measurements
S-parameter
measurements were made with a VNA that actuated the piezoelectric
were made with a VNA that actuated the piezoelectric transducers with sinusoidal waveforms transducers with sinusoidal at
waveforms at an output power of +10 dBm. This translates to a peak-to-peak
an output power of +10 dBm. This translates to a peak-to-peak voltage of just two volts across a 50 Ωvoltage of just two volts
across
load anda 50four
Ω load
voltsand fouravolts
across high across
impedancea highload,
impedance
such as load,
a PZTsuch as a PZTIn
transducer. transducer.
both cases,Inthisboth is
cases, this is a dramatic reduction when compared to the several tens of volts
a dramatic reduction when compared to the several tens of volts that are used in typical time-domain that are used in typical
time-domain
ultrasonic SHM ultrasonic
systems. The SHMtime-domain
systems. The time-domain
waveforms that were waveforms
subsequently thatgenerated
were subsequently
using these
generated using these S-parameters were compared to oscilloscope
S-parameters were compared to oscilloscope measurements and were in good agreement measurements and were in good
with them
agreement with them while displaying higher signal-to-noise ratio and
while displaying higher signal-to-noise ratio and larger dynamic range. To better understand how larger dynamic range. To
better understand how the VNA measurements achieve better SNR
the VNA measurements achieve better SNR and dynamic range while actuating the PZT transducers and dynamic range while
actuating the PZT
at dramatically transducers
lower voltages,atthedramatically
architecturelower voltages, of
and operation thea architecture
VNA is briefly anddescribed
operationinofthe a
VNA is briefly
sub-section described in the sub-section below.
below.

(a) Time-Domain Measurements

In Software
Structure Time-domain based
h(t) detection and
localization algorithms
y(t)
v(t)

(b) Processing in Software

S-Parameter Time-domain based

FFT Structure Model IFFT detection and
Sx(f) localization algorithms
y(t)

v(t)

VNA
Measurements

Figure 2.
Figure Flow diagrams
2. Flow diagrams for
for (a)
(a) a time-domain-based
time-domain-based structural
structural health
health monitoring
monitoring (SHM)
(SHM) approach
approach
and an (b) S-parameter-based SHM
and an (b) S-parameter-based SHM approach.

2.1. Vector Network Analyzers

2.1. Vector Network Analyzers
A VNA measures the S-parameters of a device under test (DUT) by directly measuring the
A VNA measures the S-parameters of a device under test (DUT) by directly measuring the
component signals shown in Figure 1 in a vector (both magnitude and phase) fashion. Once these
component signals shown in Figure 1 in a vector (both magnitude and phase) fashion. Once these
measurements have been taken, they are subsequently used to calculate the S-parameters according to
measurements have been taken, they are subsequently used to calculate the S-parameters according
the expressions shown in Figure 1. The sub-systems required to make these measurements in a typical
to the expressions shown in Figure 1. The sub-systems required to make these measurements in a
VNA system are a signal source, signal separation hardware, measurement receivers and a processing
typical VNA system are a signal source, signal separation hardware, measurement receivers and a
backend. The architecture of a two-port VNA is shown in Figure 3 and the component sub-systems are
processing backend. The architecture of a two-port VNA is shown in Figure 3 and the component
described in the following sub-sections.
sub-systems are described in the following sub-sections.

2.1.1. Signal Source

The source provides the stimulus signal that is injected into the DUT to initiate the
characterization. The signal is sinusoidal, and it can excite the DUT at a single frequency or more
Aerospace 2020, 7, 33 4 of 15

2.1.1. Signal Source

Aerospace 2020, 7, x FOR PEER REVIEW 4 of 15
The source provides the stimulus signal that is injected into the DUT to initiate the characterization.
The
commonly is
signal sinusoidal,
across and of
a sequence it can excite the DUT
predetermined at a single
frequencies frequency
in the or more
bandwidth commonly
of interest. Sinceacross
VNA
aports
sequence of predetermined
are usually stimulated frequencies
sequentially,insources
the bandwidth of interest.
often include Sinceswitch
a transfer VNA ports
such are usually
as the one
stimulated sequentially, sources often include a transfer switch such as the one shown
shown in Figure 3 to enable the use of the same source at both ports and, more generally, each port in Figure 3
to enable the use of the
in an N-port network analyzer. same source at both ports and, more generally, each port in an N-port
network analyzer.
2.1.2. Signal separation hardware
2.1.2. Signal Separation Hardware
At each port, the reference and reflected signals travel on the same transmission line but they
need to each
At port, the independently
be measured reference and reflected signals
to determine thetravel on the same
S-parameters. transmissionof
A characteristic line
thebut they
signals
need to be
that can be measured independently
used to separate to determine
them is their direction the S-parameters.
of travel along theA characteristic
transmission of Directional
line. the signals
that can be used to separate them is their direction of travel along the transmission line.
elements such as directional couplers, directional bridges, circulators, and Wilkinson dividers can be Directional
elements suchpurpose.
used for this as directional
In VNA couplers,
systems,directional
directionalbridges,
couplers circulators, and Wilkinson
are more commonly useddividers
than thecan be
other
used for this purpose. In
kinds of directional elements.VNA systems, directional couplers are more commonly used than the other
kinds of directional elements.

Source

Port 1 Port 2

Directional Couplers

Z0 Z0
Device
Under
Test

Reference Reference
Receiver Measurement Measurement Receiver
Receiver Receiver

Calibration and Processing

Figure 3. Typical vector network analyzer (VNA) architecture.

Figure 3. Typical vector network analyzer (VNA) architecture.
Figure 4 shows the structure of a generic four port directional coupler. Ports 1 and 2 are directly
Figureand
connected 4 shows
form thethe structure
main lineofofathe generic fourand
coupler port directional
the coupler.
signal travels Portsthem
between 1 andwith
2 areminimal
directly
connected
loss. Portsand
3 andform thealso
4 are main line ofconnected
directly the coupler toand
form the signal
the travels
auxiliary linebetween them with
of the coupler. Theminimal
signal
loss. Ports 3 and 4 are also directly connected to form the auxiliary line of the
also travels along this line with minimal loss. An electromagnetic or electrical coupling mechanism coupler. The signal also
travels along
transfers this line fraction
a well-defined with minimal loss. power
of the signal An electromagnetic
from the main line or electrical coupling
to the auxiliary line.mechanism
The signal
transfers a well-defined fraction of the signal power from the main
transferred to the auxiliary line can then be used in another circuit, such as a measurement line to the auxiliary line. The
circuit,
signal transferred to the auxiliary line can then be used in another circuit, such
with minimal disruption to the signal on the main line. To give the coupler its directional properties, as a measurement
circuit,
this with minimal
coupling mechanism disruption
works in tosuch
the signal
a way on the
that main line.
a signal To give
traveling the direction
in one coupler its (fordirectional
example,
properties, this coupling mechanism works in such a way that a signal traveling in
Port 1 to Port 2) couples much more strongly to the auxiliary line than a signal travelling in the opposite one direction (for
example, Port 1 to Port 2) couples much more strongly to the auxiliary line than a signal
direction (Port 2 to Port 1). The key performance characteristic of a directional coupler is its directivity. travelling in
the opposite
Directivity is adirection
measure(Port of how2 tomuch
Port more
1). The key performance
strongly characteristic
a signal travelling of a directional
in the desired direction coupler
couples
is its directivity. Directivity is a measure of how much more strongly a signal
to the auxiliary line than the unwanted signal travelling in the reverse direction. Directivity travelling in theisdesired
one of
direction couples to the auxiliary line than the unwanted
the major factors affecting the accuracy of a VNA’s measurements [24–26]. signal travelling in the reverse direction.
Directivity is one of the major factors affecting the accuracy of a VNA’s measurements [24–26].
Aerospace 2020, 7, x FOR PEER REVIEW 5 of 15

Aerospace 2020, 7, 33
x FOR PEER REVIEW 55of
of 15
Port 1 Port 2
Input Port Output Port

Port 1 Port 2
Input Port Output Port

Coupling
Mechanism

Port 3 Coupling Port 4

Coupled Port Isolated Port
Mechanism

Port 3 Port 4
Coupled Port Isolated Port
Figure 4. Structure of a directional coupler.
Figure 4. Structure of a directional coupler.
2.1.3. Measurement receivers Figure 4. Structure of a directional coupler.
2.1.3. Measurement Receivers
2.1.3.These receivers receivers
Measurement
These
measure the magnitude and phase of the reference and reflection signals after
receivers measure the magnitude and phase of the reference and reflection signals after
the signal separation stage. A typical measurement receiver consists of an analog part and a digital
the signal separation stage. Athe typical measurement receiver consists of an analog part and a digital
part. These receivers
The analog measure
portion magnitude
of the receiver is most andwidely
phase of the reference
implemented as aand reflection
heterodyne signals
receiver after
that
part.
the The
signal analog
separationportion
stage.of Athe receiver
typical is most
measurement widely implemented
receiver consists ofas
ana heterodyne
analog part receiver
and a that
digital
mixes the input signal with a local oscillator (LO) signal to yield a signal at a much lower intermediate
mixesThe
part. the analog
input signal with
portion ofathe
local oscillator
receiver (LO) widely
is most signal toimplemented
yield a signal at aa heterodyne
much lower receiver
intermediate
frequency (IF). Figure 5 shows such a heterodyne receiver architecture.asThis receiver architecture that
is
frequency
mixes the (IF). Figure
input signal 5 shows
with afilteringsuch a
local oscillator heterodyne receiver architecture. This receiver architecture
favored because subsequent at the IF(LO) signal to
frequency yield in
results a signal at a much
a smaller receiver lower intermediate
bandwidth, and
is favored (IF).
frequency because subsequent
Figure 5 minimizing
shows such filtering at the IF frequencyarchitecture.
a heterodyne results in a smaller receiver bandwidth,
this has the advantage of the integrated receiver
broadband noise which This receiver
in turn architecture
improves receiveris
and this has
favored because the advantage
subsequent of minimizing
filtering the integrated broadband noise which in turn improves
sensitivity and dynamic range. TheatIFthefilter
IF frequency
also doublesresultsupin as
a smaller receiver bandwidth,
an anti-aliasing filter for andthe
receiver
this has sensitivity
the advantage andof dynamic
minimizing range.
the The IF filterbroadband
integrated also doubles upwhich
noise as an in
anti-aliasing
turn improvesfilterreceiver
for the
analog-to-digital converter (ADC), and sampling and conversion at the lower IF also relaxes the
analog-to-digital
sensitivity converter range.
and dynamic (ADC), The and sampling
IF filter and conversion
also doubles upat the lower
as an IF also relaxes
anti-aliasing thefor
design
design requirements of the ADC, especially in microwave applications. Once the signal filter
is converted the
requirements
analog-to-digital of the ADC, especially in microwave applications. Once the signal is converted to digital
to digital form, theconverter (ADC),
digital signal and sampling
processing and conversion
(DSP) portion at thedoes
of the receiver lower IF also
further relaxes the
processing to
form, the
design digital signal
requirements of processing
the ADC, (DSP) portion
especially in of the receiver
microwave does further
applications. Onceprocessing
the signaltoismeasure
convertedthe
measure the magnitude and phase of the signal as required.
magnitude
to and the
digital form, phase of the
digital signal
signal as required.
processing (DSP) portion of the receiver does further processing to
measure the magnitude and phase of the signal as required.
LNA Mixer IF Filter

LNA Mixer IF Filter

Input Signal ADC DSP

Input Signal ADC DSP

LO
Figure 5. Block diagram
LO of a heterodyne receiver.
Figure 5. Block diagram of a heterodyne receiver.
2.1.4. Processing Backend
2.1.4. Processing backend Figure 5. Block diagram of a heterodyne receiver.
Once the magnitude and phase measurements of the signals are taken, they are further processed
to yield
Once
2.1.4. thethe
S-parameters
Processing backendand
magnitude andto format
phase them into representations
measurements of the signalsthat areare easy to usethey
taken, and are
interpret
furtherfor
a wide variety
processed of applications.
to yield the S-parameters Examples
and to of thesethem
format are magnitude-phase
into representations plots,
thatreal-imaginary
are easy to useplots,
and
Smith Once
interpret thepolar
charts,
for a magnitude
wide plots,
varietyandofsweeps,
linear phase measurements
logarithmic
applications. of the
sweeps
Examples signals
ofand are
otherare
these such taken,
formats.they
magnitude-phaseIn are further
addition
plots,to
processed to yield
this, the processing
real-imaginary the
plots, S-parameters
backend
Smith and toplots,
also performs
charts, polar format themsweeps,
errorlinear
correctingintofunctions
representations
such as
logarithmic that areand
easyother
averaging,
sweeps to usesuch
and
calibration,
interpret
formats. Infor a wide
embedding/de-embedding,
addition variety
to this, power
the of correction,
applications.
processing andExamples
backend other
also such of these
errorare
functions.
performs magnitude-phase
correcting functions such plots,
as
real-imaginary plots, Smith charts, polar plots, linear sweeps, logarithmic
averaging, calibration, embedding/de-embedding, power correction, and other such functions. sweeps and other such
formats. In addition
2.2. Attaining Higher toSNR this, the processing backend also performs error correcting functions such as
averaging,
2.2. Attaining calibration,
Higher SNR embedding/de-embedding, power correction, and other such functions.
The sensitivity of a measurement system is the minimum signal level that it can receive with
acceptable quality. Sensitivity
The sensitivity is determined byisthe
thethermal noise floor,level
the measurement bandwidth,
2.2. Attaining Higherof SNRa measurement system minimum signal that it can receive with
the additional
acceptable noiseSensitivity
quality. introducedisby the system,
determined byand
thethe minimum
thermal noiseSNRfloor,required by the system.
the measurement At room
bandwidth,
the The sensitivity
temperature
additional (298
noise of a measurement
K),introduced
this relationship system
by theissystem,
captured is
andtheEquation
in minimum
the minimum(1),signal
asSNR level
shown that itby
below
required can
thereceive
[27]: system.with
At
acceptable
room quality.(298
temperature Sensitivity
K), thisisrelationship
determinedisby the thermal
captured noise floor,
in Equation (1), the measurement
as shown bandwidth,
below [27]:
the additional noise introduced Pmin by= −174 dBm + NF
the system, and+the minimum
10log(B) + SNR SNR
min , required by the system. At (1)
P = −174 dBm + NF + 10log(B) + SNR
room temperature (298 K), this relationship is captured in Equation (1), as shown below [27]:
min min , (1)

Pmin = −174 dBm + NF + 10log(B) + SNRmin, (1)

Aerospace 2020, 7, x FOR PEER REVIEW 6 of 15

Aerospace 2020, 7, 33
In this equation, Pmin is the minimum acceptable signal power in dBm, −174 dBm is the thermal noise 6 of 15

floor (a function of the Boltzmann constant and temperature), NF is the system’s noise figure in
decibels (dB), B isPthe measurement bandwidth in hertz (Hz) and SNRmin is the system’s minimum
In this equation, min is the minimum acceptable signal power in dBm, −174 dBm is the thermal
SNR requirement in dB. Taken together,
noise floor (a function of the Boltzmann constant the firstand
three terms on the
temperature), NF right define
is the the system’s
system’s noise
noise figure in
floor.
decibels (dB), B is the measurement bandwidth in hertz (Hz) and SNRmin is the system’s minimum
SNR Once the minimum
requirement SNR requirement
in dB. Taken together, the is fixed,
first Equation
three terms on (1)the
says that
right therethe
define aresystem’s
two ways in which
noise floor.
the minimum acceptable signal level can be lowered. The first method is to reduce
Once the minimum SNR requirement is fixed, Equation (1) says that there are two ways in which the system’s noise
figure
the by improving
minimum acceptable the system’s
signal levelnoise
canperformance.
be lowered. This comes
The first at the iscost
method of increased
to reduce power
the system’s
consumption, a larger area, increased design effort, and several other factors.
noise figure by improving the system’s noise performance. This comes at the cost of increased power The second way to
improve sensitivity is to reduce the bandwidth over which the measurement
consumption, a larger area, increased design effort, and several other factors. The second way to is taken. As illustrated
in Figuresensitivity
improve 6, reducing is tothe bandwidth
reduce reducesover
the bandwidth the which
broadband noise that isiscaptured
the measurement taken. Asalongside
illustratedthe in
signal that we are interested in measuring and this reduction in the captured
Figure 6, reducing the bandwidth reduces the broadband noise that is captured alongside the signal noise results in higher
that
SNR.
we areLowering
interestedthe measurement
in measuring and bandwidth
this reductioncomes atcaptured
in the the cost of increasing
noise results inthe time SNR.
higher over which
Lowering the
measurement is taken. The two quantities are inversely related through
the measurement bandwidth comes at the cost of increasing the time over which the measurement is the time-frequency
uncertainty
taken. principle.
The two quantities are inversely related through the time-frequency uncertainty principle.

Bandwidth

Signal Power
Power

Signal
Noise
PSD
IF Filter Noise Power

Frequency

Reduced
Bandwidth
Power

Signal Power

Noise Power
Frequency

Illustrationshowing
Figure6.6.Illustration
Figure showingthe
theeffect
effectof
of reducing
reducing filter
filter bandwidth
bandwidth on
on the
thesignal-to-noise
signal-to-noiseratio
ratio(SNR).
(SNR).

VNAs exploit the second method to a achieve high sensitivity. The user can vary the measurement
VNAs exploit the second method to a achieve high sensitivity. The user can vary the
bandwidth by adjusting the intermediate frequency filter bandwidth (IFBW). Taking a case where
measurement bandwidth by adjusting the intermediate frequency filter bandwidth (IFBW). Taking a
the IFBW is reduced by a factor of N, Equations (2) to (4) below show the expected improvement
case where the IFBW is reduced by a factor of N, Equations (2) to (4) below show the expected
in sensitivity.
improvement in sensitivity.
Pmin,new = −174 dBm + NF + 10log(Bnew ) + SNRmin , (2)
Pmin,new = −174 dBm + NF + 10log(Bnew) + SNRmin, (2)
Pmin,new = −174 dBm + NF + 10log(B0 /N) + SNRmin , (3)

Pmin,new P=min,new
−174=dBm
−174+dBm + NF
NF + + 10log(B
10log(B 0/N) + SNRmin,
0 ) + SNRmin − 10log(N),
(3)
(4)

It is clear that to satisfy the= same

Pmin,new SNR +requirement,
−174 dBm NF + 10log(BP0min,new
) + SNRcan be10log(N),
min − up to 10log(N) times smaller(4)
than the power of the minimum acceptable signal at the initial measurement bandwidth, B0 .
√ requirement, Pmin,new can be up to 10log(N) times smallerIf the
It is clear that to satisfy the same SNR
signal is sinusoidal, this translates to a N smaller amplitude and this benefit comes at the cost of
than the power of the minimum acceptable signal at the initial measurement bandwidth, B0. If the
an N-fold increase in the measurement time. It is worth noting that the VNA measurements are made
signal is sinusoidal, this translates to a √ smaller amplitude and this benefit comes at the cost of an
in the frequency domain, hence this sensitivity improvement manifests itself as a better measurement
Aerospace 2020, 7, x FOR PEER REVIEW 7 of 15

N-fold
Aerospaceincrease
2020, 7, 33in the measurement time. It is worth noting that the VNA measurements are made 7 of 15
in the frequency domain, hence this sensitivity improvement manifests itself as a better measurement
of the S-parameters. However, through Parseval’s theorem, the total noise energy of a signal in the
of the S-parameters.
frequency domain is equalHowever,to thethrough Parseval’s
total noise energytheorem,
of that samethe total
signal noise energy
in the time of a signalHence,
domain. in the
frequency domain is equal to the total noise energy of that same signal
the higher quality spectrum measurements at a smaller IFBW translate to higher quality time-domain in the time domain. Hence,
the higher quality
waveforms when the spectrum
inversemeasurements
Fourier transform at a smaller
is appliedIFBW translate
to the spectrum. to higher quality time-domain
waveforms when the inverse Fourier transform is applied to the spectrum.
The principle described above is what was exploited in prior work to make low-voltage
The principle
VNA-based SHM described
measurementsabove isthatwhathave
was exploited
much better in prior
SNRwork to make low-voltage
performance VNA-based
than time-domain
SHM measurements
measurements [14]. Inthat have
that much
work, thebetter SNR performance
following VNA parameters than were
time-domain measurements
identified as the key ones [14].
In that work, the following VNA parameters were identified as the key
affecting the duration and quality of the generated time-domain signals: the VNA frequency step, ones affecting the duration
and quality
VNA starting of the generated
frequency, VNAtime-domain signals: theand
ending frequency, VNAthe frequency step, VNA starting
VNA intermediate frequency,
frequency filter
VNA ending
bandwidth frequency, and the VNA intermediate frequency filter bandwidth (IFBW).
(IFBW).
Before designing
Before designingand andimplementing
implementingSHM SHMhardware
hardwarebasedbasedon onS-parameter
S-parametermeasurements,
measurements,ititisis
importantto
important todemonstrate
demonstrate the the utility
utility of
of these
thesemeasurements
measurements in in detecting
detecting and and localizing
localizing defects.
defects. In
Inthe
the
next section,
next section, wewe describe and reportreport thethe results
resultsofofdefect
defectlocalization
localizationexperiments
experimentsthat thatwe
weconducted
conducted in
widely used aerospace materials.
in widely used aerospace materials.

3.Materials
3. Materialsand
andMethods
Methods
Experiments to
Experiments tovalidate
validatethethe
S-Parameter
S-Parameterdamage localization
damage method
localization were carried
method were out in aluminum
carried out in
and carbon-fiber-reinforced
aluminum polymer (CFRP)
and carbon-fiber-reinforced plates.
polymer These plates.
(CFRP) test structures are shown
These test in Figure
structures 7 and we
are shown in
provide details on how they were prepared and how the measurements were
Figure 7 and we provide details on how they were prepared and how the measurements were taken taken in Appendix A.
After taking the measurements, the S-parameters for each transducer-to-transducer
in Appendix A. After taking the measurements, the S-parameters for each transducer-to-transducer path are extended
to have
path are aextended
starting tofrequency of 0 Hzfrequency
have a starting by zero-padding.
of 0 Hz byThey are then combined
zero-padding. They are with a digitally
then combined
generated input tone-burst waveform, v(t), to generate time-domain signals for
with a digitally generated input tone-burst waveform, v(t), to generate time-domain signals for defectdefect localization
using the procedure
localization using the described
procedure below and illustrated
described below andin Figure 8. First,
illustrated inthe Discrete
Figure Fourier
8. First, theTransform
Discrete
(DFT), V(f), of the input tone-burst was taken. For each transducer-to-transducer
Fourier Transform (DFT), V(f), of the input tone-burst was taken. For each transducer-to-transducer path, the frequency
spectrum
path, of the pitch-catch
the frequency spectrum output
of thesignal, Y(f), was
pitch-catch obtained
output signal,by multiplying
Y(f), V(f) and
was obtained the path’s SV(f)
by multiplying 21 (f).
Finally,
and the the time-domain
path’s pitch-catch
S21(f). Finally, signal, y(t),pitch-catch
the time-domain was generated by y(t),
signal, taking thegenerated
was Inverse Discrete
by taking Fourier
the
Transform (IDFT) of Y(f). These time-domain pitch-catch signals then served as
Inverse Discrete Fourier Transform (IDFT) of Y(f). These time-domain pitch-catch signals then servedinputs to the widely
used
as reconstruction
inputs to the widelyalgorithm for probabilistic
used reconstruction inspection
algorithm of defectsinspection
for probabilistic (RAPID), of which
defectswe(RAPID),
used for
damage
which welocalization
used for damage[18]. localization [18].

7
8 6
1 5
2
3 4

(a) (b) (c)

7
8 6
1 5
2
3 4

(d) (e)
Figure
Figure7.7.(a)
(a)Measurement
Measurementsetup, (b)(b)
setup, aluminum
aluminum defect Location
defect Location1, (c)
1, aluminum
(c) aluminumdefect Location
defect 2, (d)2,
Location
carbon-fiber-reinforced polymer
(d) carbon-fiber-reinforced (CFRP)
polymer defect
(CFRP) Location
defect 1, and
Location (e) CFRP
1, and (e) CFRPdefect Location
defect 2. 2.
Location
Aerospace 2020, 7, 33 8 of 15
Aerospace 2020, 7, x FOR PEER REVIEW 8 of 15

VNA
Measurements

S21(f) IFFT

y(t)

FFT Y(f)

v(t)
V(f)

Figure 8.
Figure Procedure for
8. Procedure for generating
generating time-domain
time-domain signals
signals from
from S-parameter
S-parameter measurements.
measurements.
4. Results
4. Results
The results of the S-parameter measurements taken in both aluminum and composite materials are
The results of the S-parameter measurements taken in both aluminum and composite materials
presented in this section. In both materials, the measured S21 spectra for select transducer-to-transducer
are presented in this section. In both materials, the measured S21 spectra for select
paths are presented followed by the time-domain waveforms generated from the spectra. After this,
transducer-to-transducer paths are presented followed by the time-domain waveforms generated
the defect localization maps generated from the time domain waveforms at a number of different
from the spectra. After this, the defect localization maps generated from the time domain waveforms
frequencies are shown.
at a number of different frequencies are shown.
4.1. Results of the Aluminum Test Structure
4.1. Results of the Aluminum Test Structure
Figure 9 shows the S21 parameter plots for the paths between PZT 1 and PZT 8 as well as PZT 4 and
Figure
PZT 8 (baseline 9 shows the S21 parameter
and defective cases withplots for theatpaths
the defect Locationbetween PZT paths
1). These 1 andwere
PZT chosen
8 as well to as PZT 4
compare
and PZT 8 (baseline and defective cases with the defect at Location
the case of not passing through the defect versus passing directly through the defect. As expected, 1). These paths were chosen to
compare the case of not passing through the defect versus passing
there are more pronounced differences in the baseline and defect spectra for the latter case. directly through the defect. As
expected, there aresome
In particular, moreattenuation
pronounced differences
is observed in in
thethe baseline
defect stateand defect spectra
S21 magnitude for thefor
spectrum latter
thecase.
PZT
In particular, some attenuation is observed in the defect state S 21 magnitude spectrum for the
4 to PZT 8 path across all frequencies, but especially at frequencies above 700 kHz. This attenuation is
PZT
likely4dueto PZT
to the8 absorption
path across of all
somefrequencies,
of the Lamb butwaves’
especiallyenergy at by
frequencies
the sealantabove
material700 that
kHz. This
acts as
attenuation is likely due to the absorption of some of the Lamb waves’ energy
a defect in our experiments. Furthermore, there is greater deviation across a wider range of frequencies by the sealant material
that
in theacts as aspectra
phase defect ofin the
ourPZTexperiments.
4 to PZT 8Furthermore,
path. We suspect therethatis greater deviation
this greater across
deviation cana be
wider range
attributed
of
to frequencies
a greater degree in theofphase spectrabetween
interaction of the PZT the4defect
to PZT and8 path. We suspect
the Lamb wavesthat this greater
travelling on this deviation
path as
can be attributed to a greater degree of interaction between the defect
compared to the more distant PZT 1 to PZT 8 path. Finally, since the phase spectra are approximately and the Lamb waves travelling
on thisand
linear paththe as phase
compared to theunder
spectrum more distant
the defect PZTcondition
1 to PZT for 8 path. Finally,
the PZT 4 tosince
PZT the phase
8 path liesspectra
under theare
approximately linear and we
baseline phase spectrum, the expect
phase spectrum
to see slightlyunder morethe delay
defectin condition for the PZT
the time-domain 4 to PZT for
waveforms 8 path
the
lies under the baseline phase spectrum, we
defect case and this is confirmed in the waveforms in Figure 10. expect to see slightly more delay in the time-domain
waveforms for the defect
The time-domain case and generated
waveforms this is confirmed
from S21 infor
thedifferent
waveforms in Figure
paths, 10. the PZT 1 to PZT
including
The time-domain waveforms generated from S 21 for different paths, including the PZT 1 to PZT
8 and PZT 4 to PZT 8 paths, are shown in Figure 10. The digitally generated input tone-burst has
8anand PZT 4 toofPZT
amplitude 20 V8 and
paths, are shown
is centered atin300Figure
kHz. 10. TheThe digitally
defect was at generated
Locationinput
1. Eachtone-burst
plot shows hasthe
an
amplitude
first 100 µs of of 20
theV4and is centered
ms record at 300
length. kHz. Thetodefect
In addition was atvoltages
the output Locationin1.theEach plot shows
baseline the first
and defective
100 μs of the 4 ms record length. In addition to the output voltages in the
states, each plot shows the scatter signal, which is obtained by baseline subtraction. The amplitudes baseline and defective states,
of
each
the output waveforms range from 11.2 mV to 29.6 mV, corresponding to attenuations of ~100 of
plot shows the scatter signal, which is obtained by baseline subtraction. The amplitudes the
times
output
(40 waveforms
dB) relative range
to the input. from 11.2 mV to 29.6 mV, corresponding to attenuations of ~100 times
(40 dB) relative to the input.
Aerospace 2020, 7, x FOR PEER REVIEW 9 of 15

Aerospace
Aerospace2020, 7, 7,
2020, 33x FOR PEER REVIEW 99ofof1515

S21 Magnitude [dB]

S21 Magnitude [dB]

S21 Magnitude [dB]

S21 Magnitude [dB]

S21 Phase [Radians]

S21 Phase [Radians]

S21 Phase [Radians]

S21 Phase [Radians]

(b)
(a)
(b)
(a)
Figure 9. Aluminum S21 spectra for (a) Path 1-to-8 and (b) Path 4-to-8.
Figure 9. Aluminum S21 spectra for (a) Path 1-to-8 and (b) Path 4-to-8.
Figure 9. Aluminum S21 spectra for (a) Path 1-to-8 and (b) Path 4-to-8.
Baseline
With Defect at Location 1
Output Voltage for Path 1-to-8 10 -4 Scatter Signal for Path 1-to-8
4
Voltage [V]

Voltage [V]
0.02 Baseline
2
With Defect at Location 1
0 0
Output Voltage for Path 1-to-8 -2 10 -4 Scatter Signal for Path 1-to-8
-0.02
-4 4
Voltage [V]

Voltage [V]
0.02
0 0.2 0.4 0.6 0.8 1 02 0.2 0.4 0.6 0.8 1
0 0
Time [s] 10 -4 -2
Time [s] 10 -4
-0.02 -4
Output Voltage for Path 3-to-8 -4 10 Scatter Signal for Path 3-to-8
0.02
0 0.2 0.4 0.6 0.8 1 4 0 0.2 0.4 0.6 0.8 1
Voltage [V]

Voltage [V]

2
Time [s] 10 -4 Time [s] 10 -4
0 0
Output Voltage for Path 3-to-8 -4 Scatter Signal for Path 3-to-8
-2 10
0.02
-4 4
Voltage [V]

Voltage [V]

-0.02 2
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
0 0
Time [s] 10 -4 -2 Time [s] 10 -4
Output Voltage for Path 4-to-8 -4 10 -3 Scatter Signal for Path 4-to-8
-0.02
0.02 5
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Voltage [V]

Voltage [V]

Time [s] 10 -4 Time [s] 10 -4

0 0
Output Voltage for Path 4-to-8 10 -3 Scatter Signal for Path 4-to-8
0.02 5
Voltage [V]

Voltage [V]

-0.02 -5
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
0 0
Time [s] 10 -4 Time [s] 10 -4
Output Voltage for Path 3-to-5 -3 Scatter Signal for Path 3-to-5
-0.02 -5 10
0.01 0 0.2 0.4 0.6 0.8 1 2 0 0.2 0.4 0.6 0.8 1
Voltage [V]

Voltage [V]

-4
Time [s] 10 Time [s] 10 -4
0 0 -3
Output Voltage for Path 3-to-5 10 Scatter Signal for Path 3-to-5
2
Voltage [V]

Voltage [V]

0.01
-0.01 -2
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
0 0
Time [s] 10 -4 Time [s] 10 -4
-0.01 -2
0 0.2 0.4 0.6 0.8 1 0 0.2
Figure 10. Output voltage
Time [s]
and scatter
10 -4
signals at 300 kHz0.4in 0.6
aluminum.
Time [s]
0.8
10 -4
1

Figure 10. Output voltage and scatter signals at 300 kHz in aluminum.
For Paths 1-to-8 and 3-to-8, which do not pass through the defect, the waveforms for the baseline
and defective 1-to-8Figure
For Pathsconditions andare10.almost
Output
3-to-8, which voltage
do not
identical and scatter
pass
and signals
through
they yield theat defect,
300signals
scatter kHzthe
in aluminum.
waveforms
with maximum for the baseline
amplitudes
and defective conditions are almost identical and they yield
of about 0.5 mV. For Paths 3-to-5 and 4-to-8, the outputs for the defective case show significantscatter signals with maximum
For Paths
amplitudes of 1-to-8 0.5
andmV.3-to-8,
ForSwhich do not pass through the outputs
defect, the
forwaveforms for the
casebaseline
attenuation, as about
predicted by the Paths
21 3-to-5
amplitude and 4-to-8,
spectra. Asthe
a result, with the defective
maximum amplitudes showof
and defective
significant conditions
attenuation, are almost
as predicted by identical
the and they spectra.
S21 are
amplitude yield scattera signals with maximum
maximum
2.5 mV and 4.7 mV, respectively, the scatter signals almost an order ofAs result, larger
magnitude with than those
amplitudesofof2.5
amplitudes about and
0.5 mV. For Paths 3-to-5 and 4-to-8, the outputs for the defective case show
of Paths 1-to-8 andmV 3-to-8. 4.7 mV, respectively, the scatter signals are almost an order of magnitude
significant
larger than 11 attenuation,
those of Paths as predicted
1-to-8 and 3-to-8.by the S21 amplitude spectra. As a result, with maximum
Figure shows damage localization results for the two aluminum defects for a generated input
amplitudes
Figuresignalof 2.5
11 shows mV and
damage 4.7 mV, respectively,
localization the scattertwosignals are almost an
fororder of magnitude
tone-burst centered at 300 kHz. Asresults for the
the images show, aluminum defects
there is good agreementa generated
between input
the
larger
tone-burstthan those
signal of Paths
centered 1-to-8
at 300 and
kHz. 3-to-8.
As the images show, there is good agreement between the
physical defect locations and sizes and the damage localization maps generated using the RAPID
Figure
physical defect 11 shows damage
locationsthese localization
andimages,
sizes and results for the two aluminum defects for a generated input
algorithm. In generating thethe
valuedamage
of β, the localization
parameter maps generated
that controls using
the size theeffective
of the RAPID
tone-burstIn
algorithm. signal centered
generating at 300
these kHz. the
images, As value
the images
of β, show,
the there is good
parameter that agreement
controls the between
size of thethe
elliptical distribution area, was set to 1.05 in line with the guidance on preserving a good resolution
physical
effective defect locations
ellipticalartifacts and
distribution sizes
area,givenand
was in the
setthedamage
to work
1.05 in localization
line maps
with thethe generated
guidance using the RAPID
while minimizing that was that pioneered methodon preserving
[18]. Once they a good
have
algorithm.
resolution In generating
whilethe minimizing these images,
artifacts the value
thateasier of
was comparison β,
given in the ofthe parameter
work that controls
that pioneered the
the method size of the
[18].
been measured, S-parameters enable detection and localization results at
effective
Once they elliptical
have been distribution area, was set to 1.05 in line with the guidance on preserving a good
different frequencies thanmeasured,
the traditional the time-domain
S-parametersapproach.enable easier comparison
Waveforms of detection
at new frequencies canand
be
resolution
localization while
results minimizing
at different artifacts
frequencies that was given in
than the traditional the work that
time-domain pioneered
approach.the method
Waveforms [18].
generated from the recorded S-parameter measurements and processed for detection and localization
atOnce
purposes
they have can
new frequencies been measured,
in a matter of be generated
seconds rather
the S-parameters
from
thanthe
enable easier
recorded S-parameter
reconfiguring
comparison and
measurements
a time-domain-based
of processed
measurement
detection for
system
and
localization results at different frequencies than the traditional time-domain approach. Waveforms
to take a new set of measurements at each new frequency. Consequently, this enables the efficient
at new frequencies can be generated from the recorded S-parameter measurements and processed for
Aerospace 2020, 7, x FOR PEER REVIEW 10 of 15

detection and localization purposes in a matter of seconds rather than reconfiguring a

Aerospace 2020, 7, 33 10 of 15
time-domain-based measurement system to take a new set of measurements at each new frequency.
Consequently, this enables the efficient identification of optimal test frequencies for a wide variety of
defects acrossofdifferent
identification optimal materials. Figurefor
test frequencies 12 ashows the defect
wide variety localization
of defects acrossresults formaterials.
different the two
aluminum
Figure defect the
12 shows locations
defect at 150 kHz, 300
localization kHz,
results forand
the500
twokHz. As the defect
aluminum imageslocations
show, there is some
at 150 kHz,
variation
300 with500
kHz, and frequency
kHz. Asinthetheimages
localization
show, results,
there is and,
someinvariation
this case,with
the results at 300
frequency kHzlocalization
in the exhibit the
best correspondence
results, to the
and, in this case, the damage
results atlocations.
300 kHz exhibit the best correspondence to the damage locations.

Figure 11.
Figure 11. Defect
Defect localization results at
localization results at aluminum
aluminum Location
Location 11 with
with actuation
actuation at
at 300
300 kHz.
kHz.

(a) (b) (c)

(d) (e) (f)

Figure 12.
Figure Defect localization in aluminum at different frequencies.
12. Defect frequencies. Defect Position
Position 1 at (a) 150 kHz,
(b) 300
(b) 300 kHz,
kHz, and
and (c)
(c) 500
500 kHz,
kHz, and
and Defect
Defect Position
Position 22 at
at (d)
(d)150
150kHz,
kHz,(e)
(e)300
300kHz,
kHz,and
and(f)
(f)500
500kHz.
kHz.

4.2. Results of
4.2. Results of the
the CFRP
CFRP Test
Test Structure
Structure
The
The results
results ofof the
the CFRP
CFRP coupon
coupon largely
largely follow
follow the
the same trends as
same trends as those
those observed
observed in in aluminum.
aluminum.
Figure 13 shows
Figure 13 shows the S21the S spectra of two equidistant paths emanating from PZT 3 with
21 spectra of two equidistant paths emanating from PZT 3 with the defect at
the defect at
Location
Location 1.1.Path
Path3-to-8
3-to-8is is
farfar
fromfromthe the
defect and and
defect Path Path
3-to-63-to-6
passespasses
through the defect.
through In theseIn
the defect. spectra,
these
the patterns in magnitude attenuation and phase deviation due to defect introduction
spectra, the patterns in magnitude attenuation and phase deviation due to defect introduction are are similar to
those seen in the aluminum spectra.
similar to those seen in the aluminum spectra.
A
A notable
notable difference
differencein in the
the SS21 spectra of the CFRP relative to those measured in aluminum is the
21 spectra of the CFRP relative to those measured in aluminum is the
effect of orientation on the S magnitude.
effect of orientation on the S2121 magnitude. Figure Figure 14
14 shows
shows the
the S-parameter
S-parameter spectra
spectra of of equidistant
equidistant
paths with different angles of orientation. The angles of orientation are
paths with different angles of orientation. The angles of orientation are measured in a measured in a counterclockwise
fashion relative tofashion
counterclockwise the positive x-axis
relative of positive
to the an x–y-axis whose
x-axis of anorigin is at
x–y-axis the center
whose originofiseach
at thetest panel.
center of
As the figure shows, the angle of orientation does not introduce any discernable
each test panel. As the figure shows, the angle of orientation does not introduce any discernable difference in the
aluminum
difference in S21themagnitude
aluminumspectra. In CFRP,
S21 magnitude however,
spectra. a distinct
In CFRP, difference
however, is observed
a distinct between
difference paths
is observed
that run at the 0◦ and 90◦ orientations that align with the directions of the fiber weave of the composite
between paths that run at the 0° and 90° orientations that align with the directions of the fiber weave
plate and the 45◦ and 135◦ paths that run diagonal to the weave. At the peaks, the paths that run along
 Aerospace 2020, 7, x FOR PEER REVIEW 11 of 15

of the 2020,
Aerospace
Aerospacecomposite plate
2020,7,7,33x FOR PEERand the 45°
REVIEW and 135° paths that run diagonal to the weave. At the peaks, 11
11 of the
of1515
paths that run along the directions of the weave have peak S21 magnitudes that are comparable to the
ones
of the measured
compositeinplate
aluminum
and the(~45°−40and
dB)135°
andpaths
are atthat
leastrun
19 diagonal
dB and up to to
the29weave.
dB greater than
At the those
peaks, theof
thethedirections
paths that runof
diagonal the weave
paths
along in
thethe have material.
same
directionspeak S21 weave
of the magnitudes
This difference thatisSare
have peak comparable
testament
21 magnitudesto the toanisotropic
that the
areones measured
natureto
comparable in
ofthethe
aluminum
CFRP’s
ones (~ −40indB)
properties.
measured and are
These
aluminum at
(~ least
plots 19
alsodB)
−40 dB
showandandareup
that at to
there 29
is dB
least greater
a wide
19 dB than
up tothose
dynamic
and dBofgreater
range
29 theup
of diagonal
to 90 dB
than paths
thosein ofS21
inthe
the same
magnitude material.
diagonalin This difference
bothinmaterials
paths the sameacross is testament
the tested
material. to the anisotropic
frequencyisrange.
This difference nature
As discussed
testament of the CFRP’s properties.
earlier, anature
to the anisotropic VNA is of well
the
These
suited
CFRP’s plots
to also show
make that plots
measurements
properties. These thereacross
is a wide
also suchdynamic
show a wide
that range
dynamic
there ofrange.
is a wide updynamic
to 90 dBrangein S21 ofmagnitude in both
up to 90 dB in S21
materials
magnitude across thematerials
in both tested frequency
across the range.
testedAs discussed
frequency earlier,
range. a VNA is earlier,
As discussed well suited
a VNA to is
makewell

S21 Magnitude [dB]

measurements
suited to makeacross such a wide dynamic
such arange.
S21 Magnitude [dB]

measurements across wide dynamic range.

S21 Magnitude [dB]

S21 Magnitude [dB]

S21 Phase [Radians]

S21 Phase [Radians]

S21 Phase [Radians]

S21 Phase [Radians]

(b)
(a)

(b)
(a)
Figure 13. CFRP S21 spectra for (a) Path 3-to-8 and (b) Path 3-to-6.
Figure 13. CFRP S21 spectra for (a) Path 3-to-8 and (b) Path 3-to-6.
Figure 13. CFRP S21 spectra for (a) Path 3-to-8 and (b) Path 3-to-6.
S21 Magnitude [dB]
S21 Magnitude [dB]

(a)
(b)
Figure 14. Impact
(a) of path orientation on S21 magnitude in (a) aluminum and (b) CFRP.
(b)
Figure 14. Impact of path orientation on S21 magnitude in (a) aluminum and (b) CFRP.
A few time-domain waveforms that are generated from S21 measurements for the first defect
location
A in
fewCFRP are shownwaveforms
time-domain in Figure 15.
thatAsare
was the case in aluminum, Path 1-to-8for
and Path
first3-to-8,
Figure 14. Impact of path orientation ongenerated from
S21 magnitude S21aluminum
in (a) measurements the
and (b) CFRP. defect
which are far
location from the
in CFRP aredefect,
shownarein unaffected andwas
Figure 15. As exhibit
the very small
case in scatter signals,
aluminum, whereas
Path 1-to-8 andPath
Path3-to-6
3-to-8,
and Pathare
which 5-to-8 exhibit thedefect,
attenuation, slight delay,
andand largervery
scatter signals thatsignals,
are characteristic of
A few far from the
time-domain are unaffected
waveforms that are generatedexhibit
from S21small scatter
measurements whereas
for the Path
first defect
the presence
3-to-6 and of the
Path aresurface
5-to-8 defect
exhibit that was introduced.
the attenuation, slight
location in CFRP shown in Figure 15. As was the casedelay, and larger
in aluminum, Pathscatter signals
1-to-8 and Paththat are
3-to-8,
characteristic
which of thethe
are far from presence
defect,ofare
theunaffected
surface defect
and that wasvery
exhibit introduced.
small scatter signals, whereas Path
3-to-6 and Path 5-to-8 exhibit the attenuation, slight delay, and larger scatter signals that are
characteristic of the presence of the surface defect that was introduced.
Aerospace 2020, 7, x FOR PEER REVIEW 12 of 15

Aerospace 2020, 7, 33 12 of 15
Aerospace 2020, 7, x FOR PEER REVIEW 12 of 15

Figure 15. Output voltage and scatter signals at 300 kHz in CFRP.
CFRP.

The
The damage Figure 15.results
damage localization
localization Outputof
results voltage
of the
the twoanddefects
two scatter signals
defects at 300 on
introduced
introduced kHz
on in CFRP
the
the CFRP.
CFRP panel
panel are
are shown
shown inin
Figure 16. For Location 1, localization was successful at all tested frequencies
Figure 16. For Location 1, localization was successful at all tested frequencies with some variation in with some variation in
the
the sizeTheof
size ofdamage
the
the defect localization
defect hot spot. results
hot spot. For of the 2,
For Location
Location two
2, thedefects
the predictedintroduced
predicted defect on the CFRP
defect location
location is
is clearpanel
clear and are shown in
and corresponds
corresponds
toFigure
to the 16.
the true For Location
true location
location for 1,
for tests localization
tests atat 150
150 and was
and 500 successful
500 kHz,
kHz, with at all tested
with unclear frequencies
unclear results
results atat300 with
300kHz. some
kHz. This variation
This shows
shows the
thein
the size ofthat
difference
difference thethe
that defecttesthot
thetest spot. For
frequency
frequency can Location
canmake
make inin 2, the
the the predicted
quality of the
quality defect location
localization
of the localization is
resultsclear
andand
results the corresponds
and value
the of the
value of
to the
S-parametertrue
the S-parameter location
approach for
approach tests at
in efficiently 150 and 500
generating
in efficiently kHz, with
waveforms
generating waveformsunclear
and defectresults at
mapsmaps
and defect 300 kHz.
at several This shows
frequencies
at several the
from
frequencies
difference
afrom
single set that
a singleof setthe
of test
measurements. frequency can make in the quality of the localization results and the value of
measurements.
the S-parameter approach in efficiently generating waveforms and defect maps at several frequencies
from a single set of measurements.

(a) (b) (c)

(a) (b) (c)

(d) (e) (f)

Figure 16.
Figure 16. Defect
(d) localization
Defect localizationininCFRP
CFRPatatdifferent
different
(e)frequencies.
frequencies.Defect Position
Defect 1 at1 (a)
Position 150150
at (a)
(f) kHz, (b)
kHz,
300300
(b) kHz, andand
kHz, (c) (c)
500500
kHz, andand
kHz, Defect Position
Defect 2 at2 (d)
Position 150150
at (d) kHz (e),(e),
kHz 300300
kHz, andand
kHz, (f) (f)
500500
kHz.kHz.
Figure 16. Defect localization in CFRP at different frequencies. Defect Position 1 at (a) 150 kHz, (b)
300 kHz, and (c) 500 kHz, and Defect Position 2 at (d) 150 kHz (e), 300 kHz, and (f) 500 kHz.
Aerospace 2020, 7, 33 13 of 15

5. Discussion and Conclusions

The results presented in this work show the successful use of S-parameter measurements and
Fourier Transform techniques to perform ultrasonic guided wave RAPID defect localization in materials
that are widely used in the aerospace industry. The advantage of this approach is that actuation
voltages that are up to 100 times smaller than those used in time-domain-based SHM systems can
be applied to the PZT transducers while still yielding signals with good SNR and a wide dynamic
range. However, the S-parameters were measured using benchtop VNAs, which are very expensive
and share the same limitations with time-domain ultrasonic SHM systems in size and portability.
To address this drawback, the compatibility of the low-amplitude stimulation signals used in this
work with commercial CMOS integrated circuit fabrication technology can be exploited to build
a measurement platform of greatly reduced size and cost. Indeed, such an approach has been taken to
realize VNA-on-a-chip systems that make S-parameter measurements in microwave-based biomedical
sensors [28,29].
The next steps in our work are specifying and implementing such a system for SHM applications.
Even though the measurements that were made in this work to exploit time-frequency equivalency are
S-parameters, it is worth exploring if other frequency domain transfer functions can achieve the same
quality of results at such low voltages. One such transfer function which would be easier to measure
than the S-parameters is the voltage gain (both magnitude and phase) from the input piezoelectric
transducer to the output piezoelectric transducer. Another area that requires further investigation is
the sensitivity of this method to physical and environmental variables, such as adhesive type and
quantity, inter-sensor distance, and temperature variation. The results of these explorations will in
turn influence the requirements and construction of the integrated-circuit frequency-domain-based
SHM system that we envision.

Author Contributions: Conceptualization, G.N. and B.M.; Investigation, G.N.; Methodology, G.N.; Software,
G.N.; Supervision, B.M.; Validation, G.N.; Visualization, G.N.; Writing—original draft, G.N.; Writing—review &
editing, B.M. All authors have read and agreed to the published version of the manuscript.
Funding: The proposed work was funded in part by Acellent Technologies through funding from Air Force STTR
Contract #FA9550-18-C-0009.
Acknowledgments: The authors acknowledge the support of Acellent personnel and the contract monitor Byung
L. Lee from the Air Force Office of Scientific Research (AFOSR) in this effort.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to
publish the results.

Appendix A. Experimental Details

The aluminum plate is made of the 6061 alloy and measures 457.2 mm × 457.2 mm × 3 mm.
The ultra-strength lightweight CFRP composite plate used was sourced from McMaster-Carr (Elmhurst,
IL, USA), Product Number–8181K16, and has dimensions of 304.8 mm × 304.8 mm × 3.2 mm. On each
of the plates, eight Acellent SMLSSOP4NR 0.25-inch PZT transducers were mounted using LOCTITE
EA E-20HP epoxy. On both test coupons, the transducers are arranged in 152.4 mm diameter circles
with the one in aluminum centered at the coordinates (228.6 mm, 228.6 mm) and the one in CFRP
centered at (152.4 mm, 152.4 mm) with the origin being in the bottom left corner of the plate in each
case. Damage detection and localization was done within these circles. SMA connectors were mounted
on all the PZT transducers to provide a standard interface for the measurement of the S-parameters.
In each coupon, there are 28 unique transducer-to-transducer paths between the 8 transducers
and baseline S-Parameters were measured for each of these paths. After taking baseline measurements,
artificial defects in the form of patches of AT-200Y sealant by General Sealants (City of Industry, CA,
USA) were introduced to the structure. In aluminum, the first patch was centered at coordinates
(265 mm, 192.2mm) and measured 25 mm by 18 mm. The second patch in aluminum was centered at
coordinates (277.6 mm, 252.1 mm) and measured 23 mm × 13 mm. In CFRP, the first defect was centered
Aerospace 2020, 7, 33 14 of 15

at (185 mm, 185 mm) and measured 24 mm × 12 mm, the second was centered at (120 mm, 152.4 mm)
and measures 12 mm × 12 mm. After the introduction of each defect, a new set of S-parameter
measurements was taken in the respective structure. Figure 7 shows the measurement setup and the
defect locations in both structures.
Measurements in aluminum were made using a two-port Rohde & Schwarz (Munich, Germany)
ZNL3 VNA. The VNA output power was set to 0 dBm and this translates to an output voltage of 1.264
Vpp when the source is interfaced to a PZT transducer. The stimulus signal was swept from 5 kHz
to 1.024 MHz. This frequency range was swept in 250 Hz steps to enable an alias-free range of up to
4 ms in the generated time-domain signals. Finally, a 100 Hz IFBW was chosen to achieve good SNR
and high dynamic range. Under these conditions, it took 85.6 s to measure the S-parameters of each
transducer-to-transducer path.
The measurements in the CFRP coupon were taken using a Rohde & Schwarz ZNB4 VNA.
A higher output power of 10 dBm (4 Vpp into a PZT load) was used to compensate for the higher levels
of signal attenuation in some directions due to the anisotropicity of the CFRP material. The frequency
range was adjusted to 10 kHz to 1.024 MHz due to the higher lower frequency limit of the ZNB4.
The 250 Hz frequency steps and 100 Hz IFBW were maintained, and the measurement time reduced
slightly to 85.5 s per transducer-to-transducer path.

References
1. Giurgiutiu, V. Structural Health Monitoring with Piezoelectric Wafer Active Sensors, 2nd ed.; Elsevier: Amsterdam,
The Netherlands, 2014.
2. Giurgiutiu, V. Structural Health Monitoring of Aerospace Composites, 1st ed.; Academic Press: London, UK,
2016.
3. Lin, M.; Chang, F. The manufacture of composite structures with a built-in network of piezoceramics. Compos.
Sci. Technol. 2002, 62, 919–939. [CrossRef]
4. Wang, Y.; Qiu, L.; Luo, Y.; Ding, R. A stretchable and large-scale guided wave sensor network for aircraft
smart skin of structural health monitoring. Struct. Health Monit. 2019. [CrossRef]
5. Shen, Z.; Chen, S.; Zhang, L.; Yao, K.; Tan, C. Direct-Write Piezoelectric Ultrasonic Transducers for
Non-Destructive Testing of Metal Plates. IEEE Sens. J. 2017, 17, 3354–3361. [CrossRef]
6. Philibert, M.; Soutis, C.; Gresil, M.; Yao, K. Damage Detection in a Composite T-Joint Using Guided Lamb
Waves. Aerospace 2018, 5, 40. [CrossRef]
7. Wang, S.; Wu, W.; Shen, Y.; Liu, Y.; Jiang, S. Influence of the PZT Sensor Array Configuration on Lamb
Wave Tomography Imaging with the RAPID Algorithm for Hole and Crack Detection. Sensors 2020, 20, 860.
[CrossRef] [PubMed]
8. Zeng, L.; Huang, L.; Lin, J. Damage imaging of composite structures using multipath scattering Lamb waves.
Compos. Struct. 2019, 216, 331–339. [CrossRef]
9. Capineri, L.; Bulletti, A.; Calzolai, M.; Giannelli, P.; Francesconi, D. Arrays of Conformable Ultrasonic Lamb
Wave Transducers for Structural Health Monitoring with Real-time Electronics. Procedia Eng. 2014, 87,
1266–1269. [CrossRef]
10. Zhang, D.; Yu, P.; Beard, S.; Qing, P.; Kumar, A.; Chang, F. A new SMART sensing system for aerospace
structures. In Unmanned Systems Technology IX, 656107; Proc. SPIE 6561: Orlando, FL, USA, 2007.
11. Guo, Y.; Aquino, C.; Zhang, D.; Murmann, B. A Four-Channel, ±36 V, 780 kHz Piezo Driver Chip for Structural
Health Monitoring. In Proceedings of the 39th European Solid-State Circuits Conference, Bucharest, Romania,
16–20 September 2013; pp. 85–88.
12. Tang, X.; Zhao, H.; Mandal, S. A programmable CMOS transceiver for structural health monitoring.
In Proceedings of the 2018 IEEE Custom Integrated Circuits Conference (CICC), San Diego, CA, USA,
8–11 April 2018.
13. Tang, X.; Harley, J.; Bi, K.; Ozdemir, T.; Pawloski, M.; Mandal, S. Sparse Sensor Networks for Active Structural
Health Monitoring Using Highly Integrated CMOS Transceivers. In SPIE Smart Structures and Materials +
Nondestructive Evaluation and Health Monitoring; SPIE 10598: Denver, CO, USA, 2018.
Aerospace 2020, 7, 33 15 of 15

14. Huang, H.; Bednorz, T. Introducing S-parameters for ultrasound-based structural health monitoring. IEEE
Trans. Ultrason. Ferroelectr. Freq. Control. 2014, 61, 1856–1863. [CrossRef] [PubMed]
15. Zahedi, F.; Huang, H. Time-frequency analysis of electro-mechanical impedance (EMI) signature for
physics-based damage detections using piezoelectric wafer active sensor (PWAS). Smart Mater. Struct. 2017,
26, 055010. [CrossRef]
16. Islam, M.; Huang, H. Detecting severity of delamination in a lap joint using S-parameters. Smart Mater.
Struct. 2018, 27, 035006. [CrossRef]
17. Nyikayaramba, G.; Murmann, B. Towards On-Chip Measurement of S-Parameters for Ultrasonic
Guided-Wave SHM: Damage Localization in Aluminum Using S-Parameter Measurements. In Proceedings
of the 12th International Workshop on Structural Health Monitoring, Stanford, CA, USA, 10–12 September
2019.
18. Zhao, X.; Gao, H.; Zhang, G.; Ayhan, B.; Yan, F.; Kwan, C.; Rose, J. Active health monitoring of an aircraft
wing with embedded piezoelectric sensor/actuator network: I. Defect detection, localization and growth
monitoring. Smart Mater. Struct. 2007, 16, 1208–1217. [CrossRef]
19. Wang, C.; Rose, J.; Chang, F. A synthetic time-reversal imaging method for structural health monitoring.
Smart Mater. Struct. 2004, 13, 415–423. [CrossRef]
20. Michaels, J. Detection, localization and characterization of damage in plates with an in-situ array of spatially
distributed ultrasonic sensors. Smart Mater. Struct. 2008, 17, 035035. [CrossRef]
21. Li, G.; Chattopadhyay, A. Reference-free damage localization in time-space domain for structural health
monitoring of X-COR sandwich composites. J. Intell. Mater. Syst. Struct. 2018, 30, 371–385. [CrossRef]
22. Qiu, J.; Li, F.; Abbas, S.; Zhu, Y. A baseline-free damage detection approach based on distance compensation
of guided waves. J. Low Freq. Noise Vib. Act. Control 2018, 38, 1132–1148. [CrossRef]
23. Xu, C.; Yang, Z.; Tian, S.; Chen, X. Lamb wave inspection for composite laminates using a combined method
of sparse reconstruction and delay-and-sum. Compos. Struct. 2019, 223, 110973. [CrossRef]
24. Hiebel, M. Fundamentals of Vector Network Analysis, 1st ed.; Rohde & Schwarz: München, Germany, 2007.
25. Introduction to Network Analyzer Measurements: Fundamentals and Background. Available online:
http://download.ni.com/evaluation/rf/Introduction_to_Network_Analyzer_Measurements.pdf (accessed on
10 February 2020).
26. Vector Network Analyzer Basics. Available online: https://www.keysight.com/upload/cmc_upload/All/BTB_
Network_2005-1.pdf (accessed on 10 February 2020).
27. Razavi, B. RF Microelectronics, 2nd ed.; Prentice Hall: Upper Saddle River, NJ, USA, 2012.
28. Niitsu, K.; Nakanishi, T.; Murakami, S.; Matsunaga, M.; Kobayashi, A.; Karim, N.; Ito, J.; Ozawa, N.; Hase, T.;
Tanaka, H.; et al. A 65-nm CMOS Fully Integrated Analysis Platform Using an On-Chip Vector Network
Analyzer and a Transmission-Line-Based Detection Window for Analyzing Circulating Tumor Cell and
Exosome. IEEE Trans. Biomed. Circuits Syst. 2019, 13, 470–479. [CrossRef] [PubMed]
29. Nehring, J.; Schutz, M.; Dietz, M.; Nasr, I.; Aufinger, K.; Weigel, R.; Kissinger, D. Highly Integrated 4–32-GHz
Two-Port Vector Network Analyzers for Instrumentation and Biomedical Applications. IEEE Trans. Microw.
Theory Tech. 2017, 65, 229–244. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).