Detection, localization, and characterization of impact damage in composites using in situ transducers are important objectives for the aerospace industry to both reduce maintenance costs and prevent failures. A network of piezoelectric transducers spatially distributed over an area of interest is one practical configuration for utilizing guided waves to accomplish these objectives. Detecting and localizing barely visible impact damage with such a sparse array has been demonstrated in prior work, and improvements in localization were demonstrated by incorporating fairly crude estimates of scattering patterns in the imaging algorithms. Here we obtain more estimates of scattering patterns from a simulated defect by employing baseline subtraction of wavefield data recorded in a circle centered at the scatterer. Scattering patterns are estimated from the wavefield residual signals before and after simulated damage is introduced and the estimated scattering patterns are then incorporated into sparse array imaging via the minimum variance imaging method. Images created with different scattering patterns are compared and the efficacy of the methodology is assessed.
Periodic inspection of aerospace structures, while essential for ensuring their safety, incurs significant costs over a
structure’s life and also can result in significant loss of service. Structural health monitoring (SHM), which is also
referred to as in situ nondestructive evaluation (NDE), offers the promise of more frequent assessments of structural
integrity with little or no loss of service; however, such systems are not in common use. Here we consider a combined
SHM and NDE approach to inspection of composite, plate-like components where the SHM system detects sites of
possible damage and the follow-up NDE method utilizes the in situ SHM sensors to facilitate the inspection. The
specific SHM approach considered is that of a sparse guided wave array using simple transducers that are spatially
distributed on the structure. The NDE approach is non-contact guided wavefield imaging whereby one or more of the
SHM transducers is used as a source and full wavefield data are recorded over the area of interest. This method has the
advantage over conventional ultrasonic methods of being non-contact and requiring minimal surface preparation. Sparse
array and wavefield data from a composite specimen with simulated sites of damage are presented here to illustrate the
concept. Damage is simulated via glued-on steel plate pieces at multiple locations, and localization is performed using
delay-and-sum imaging. A small, single site of simulated damage is well-localized, whereas larger and multiple sites of
damage are not; however, their presence is readily detected. The follow-up wavefield imaging using a single sparse
array transducer as a source is able to not only locate the sites of damage, but is able to provide a reasonable estimate of
their sizes.
Guided waves are being considered for structural health monitoring (SHM) applications, and they can also be used to
reduce subsequent inspection times once defects are detected. One proposed SHM method is to use an array of
permanently attached piezoelectric transducers to generate and receive guided waves between the various transducer
pairs. The interrogation can be done on a continuous or periodic basis to assess the health of the structure. Once defects
are suspected in the structure, the traditional approach is to disassemble components for conventional nondestructive
evaluation (NDE); however, this is an expensive and time consuming process. A less expensive alternative to
conventional NDE is to record acoustic wavefield images of guided waves generated from the attached transducers.
These images clearly show details of guided waves as they propagate outward from the source, reflect from structural
discontinuities and specimen boundaries, and scatter from any damage sites within the structure. However, the recorded
waves are typically narrowband to enable effective visualization of echoes that are relatively compact in time. In this
paper, we consider wavefield images that are recorded from a chirp excitation, which offers the advantage of high
quality broadband data from a single excitation. However, responses are not directly useful because the received echoes
are too extended in time. Signals are post-processed to obtain multiple narrowband and broadband responses containing
echoes that are more compact in time to enable visualization of guided waves interacting with structural features. This
technique is demonstrated on an aluminum plate that contains attached stiffeners and glued-on piezoelectric disc
transducers. Wavefield data are recorded using an air-coupled transducer scanned over the plate surface while one of the
attached transducers acts as a guided wave source. Waves interacting with the stiffener and the inactive discs are
analyzed via broadband and narrowband processing at multiple frequencies.
KEYWORDS: Transducers, Waveguides, Signal to noise ratio, Aluminum, Data acquisition, Ultrasonics, Ferroelectric materials, Structural health monitoring, Signal processing, Transmitters
Guided wave imaging has shown great potential for structural health monitoring applications by providing a way to
visualize and characterize structural damage. For successful implementation of delay-and-sum and other elliptical
imaging algorithms employing guided ultrasonic waves, some degree of mode purity is required because echoes from
undesired modes cause imaging artifacts that obscure damage. But it is also desirable to utilize multiple modes because
different modes may exhibit increased sensitivity to different types and orientations of defects. The well-known modetuning
effect can be employed to use the same PZT transducers for generating and receiving multiple modes by exciting
the transducers with narrowband tone bursts at different frequencies. However, this process is inconvenient and timeconsuming,
particularly if extensive signal averaging is required to achieve a satisfactory signal-to-noise ratio. In
addition, both acquisition time and data storage requirements may be prohibitive if signals from many narrowband tone
burst excitations are measured. In this paper, we utilize a chirp excitation to excite PZT transducers over a broad
frequency range to acquire multi-modal data with a single transmission, which can significantly reduce both the
measurement time and the quantity of data. Each received signal from a chirp excitation is post-processed to obtain
multiple signals corresponding to different narrowband frequency ranges. Narrowband signals with the best mode purity
and echo shape are selected and then used to generate multiple images of damage in a target structure. The efficacy of
the proposed technique is demonstrated experimentally using an aluminum plate instrumented with a spatially
distributed array of piezoelectric sensors and with simulated damage.
The development of effective damage imaging and characterization tools is a challenging task because of the
dispersive and multi-modal nature of Lamb waves. An additional problem is the need for baseline data that is
required by a number of existing techniques. This paper presents the development of imaging algorithms applied
to filtered wavefield data received from piezoelectric disc sources. Frequency-wavenumber filtering is used to
separate incident/backscattered waves and individual wave modes. Filtered data are provided as input to imaging algorithms that detect damage and estimate its location. The implementation of incident and backscattered waves separation procedures avoids the need for a baseline, while mode separation permits the analysis of modes that are most sensitive to damage. The proposed algorithm is verified experimentally for damage detection on an aluminum plate.
Guided waves generated by a spatially distributed array of piezoelectric are being evaluated by many researchers for
structural health monitoring applications. These surface-mounted transducers, which are typically Lead Zirconate
Titanate (PZT), are generally assumed to be both undamaged and properly bonded to the host structure during usage.
However, this assumption may not be valid, particularly after long term operation under realistic conditions. Existing
transducer diagnosis techniques often identify PZT defects by comparing current data to baseline data previously
measured from the pristine condition of the bonded transducers. This baseline-dependent approach can result in false
alarms because of its susceptibility to operational, structural and environmental variations. In this study, a methodology
for PZT transducer diagnosis is developed to identify damaged or poorly bonded transducers by quantifying the degree
of linear reciprocity for waves propagating between pairs of surface-mounted transducers on metallic structures. The
proposed method does not require direct comparisons of signals to baselines, and also is independent of wave mode(s),
excitation signal, structural complexity and edge reflections. The efficacy of the proposed diagnostic technique is
evaluated via experiments with PZT transducers instrumented on an aluminum plate under varying environmental and structural conditions and also on a complex structure.
Full acoustic wavefield data were acquired from an aluminum plate with various structural discontinuities and artificial
defects using an air-coupled transducer mounted on a scanning stage. Piezoelectric transducers permanently mounted on
the specimen were used as wave sources. These source transducers were elements of a permanently attached sparse
array. A time series of wavefield images clearly shows details of guided waves as they propagate outward from the
source, reflect from specimen boundaries, and scatter from discontinuities within the structure. Distinct S0 and A0 Lamb
wave modes are directly visible on constant time snapshots of the captured wavefield. However, the waves propagating
outward from the source, and waves reflected from boundaries, obscure the weaker waves that are scattered from defects.
To facilitate analysis of weaker scattered waves, source waves are removed from the full wavefield data using both time
and frequency domain methods. The effectiveness of each method is evaluated in the wavenumber-wavenumber domain
and results are fused to obtain images of scattered wavefields. The method is demonstrated on a through hole, to which a
notch is added to simulate a crack. The angular dependence of the scattered wavefield is experimentally determined for
source waves incident on the notch from two directions, one toward the side of the notch and the other toward the end of
the notch.
Several imaging algorithms are being considered for localizing damage in plate-like structures by analyzing changes in
signals recorded from permanently mounted guided wave sensor arrays. Delay-and-sum type algorithms have been
shown to be effective for damage localization, but exhibit side lobes that significantly reduce the signal-to-noise ratio.
Adaptive algorithms such as MVDR (minimum variance distortionless response) can provide significant reduction in the
amplitude of side lobes. Additional improvements in image quality are possible if assumptions can be made concerning
the scattering characteristics of the damage site. In the work presented here, the efficacy of the adaptive imaging
algorithms is evaluated using both simulated and experimental waveform data. The simulated waveform data is
generated by ray tracing and incorporates edge reflections, nominal dispersion curves, and a variety of angular
scattering patterns for scatterers with cylindrical symmetry. The effect on image quality of mismatch between actual and
assumed scattering patterns is evaluated. Images generated from the simulated waveform data are compared to those
generated from experimental data for scattering from a 6 mm through-hole in an aluminum plate. The images are in
good agreement, and knowledge of scattering characteristics is shown to significantly improve imaging results.
Ultrasonic methods have been implemented for in situ sizing of fatigue cracks near fastener holes. These techniques,
however, only provide an estimate at the time of the measurement and cannot predict the remaining
life of the structure. In contrast, statistical crack propagation approaches model the expected fatigue life based
on worst-case fatigue process assumptions. The authors have recently developed a Kalman filter approach for
combining ultrasonic observations with crack growth laws. An ultrasonic angle-beam technique, combined with
an energy-based wave propagation model, serves as the measurement model. Paris's crack growth equation acts
as the system model for crack propagation. For simulated data, this approach provided more accurate crack
size estimates than either the ultrasonic measurements or crack growth approach alone. Presented here are
experimental results to assess the ability of the Kalman filter to provide reasonable crack size estimates.
Sparse ultrasonic arrays spatially distributed over a large area of a structure have been proposed and tested in the
laboratory for in situ detection and localization of damage. Detection algorithms are typically based upon comparison to
a baseline, where differences not explained by benign environmental effects are interpreted as damage. Most
localization methods are either based upon an arrival time analysis of differenced signals or spatial distribution of a
damage index. Triangulation and delay-and-sum type methods fall into the first category and, under ideal conditions,
can accurately locate discrete damage such as a single crack. Methods in the second category do not rely on precise
timing of scattered signals, but are limited in their ability to precisely locate discrete damage using a small number of
sensors. This paper evaluates the effectiveness of both types of methods for locating a single site of discrete damage,
and considers the degradation in performance resulting from errors in both wave speed and transducer locations.
Several different strategies are being considered for ultrasonic structural health monitoring systems using a variety of
approaches. Guided wave techniques for interrogating large
plate-like structures have probably generated the most
interest; these methods have the potential of monitoring large areas with a low sensor density while remaining sensitive
to defects. The acousto-ultrasonic nondestructive evaluation method has motivated the use of long-time, reverberating
waves which "fill" a structure and hence monitor large areas. Local methods based upon several different wave modes
have been considered for monitoring known "hot spots" such as fastener holes and critical bonds. Presented here are
examples of these three strategies where the purpose is to both show progress which has been made and illustrate key
issues, mainly in the context of aerospace applications. The progress and problems thus far show both the promise of
ultrasonic structural health monitoring and the significant challenges in moving from the laboratory to deployed systems.
Corrosion is detrimental to the structural integrity of many critical components, and ultrasonic methods are routinely
used in the field to make thickness measurements at points of interest. However, is often difficult to assess the true
extent of corrosion damage because of the likelihood of missing small corroded areas and the difficulty in mapping the
extent of large corroded areas without an extensive number of time consuming measurements. Guided ultrasonic waves
have the potential to both detect corrosion as early as possible and reduce the subsequent inspection time. This paper
presents results from a study using Lamb waves to quantify the area extent of corrosion in an aluminum plate specimen.
A sparse array of ultrasonic transducers was attached to an aluminum plate, and broadband excitation methods were used
to generate both symmetric and anti-symmetric Lamb wave modes. As has been demonstrated in previous studies, the
through transmission response recorded from each transmit-receive pair may be analyzed to determine if a defects exists
and approximately determine its location. This paper presents a method to determine the exact location and quantify the
extent of the corroded area using an acoustic wavefield imaging method. Lamb waves are generated using one of the
permanently attached transducers as a source, and the acoustic wavefield is captured on the surface of the plate using an
air-coupled transducer as a receiver. Full wavefield data are recorded as the receiver is scanned over the specimen
surface, and wavefield images are processed to remove the strong incident wave and enhance the weaker scattered
waves. The amplitude at the crest of the leading Lamb mode (S0) is analyzed to produce spatial images of defective
areas. Measured length and area results from these images compare very favorably with actual defect sizes. Results are
also presented for scattering from a through hole with a simulated crack.
Permanently attached piezoelectric sensors arranged in a spatially distributed array are under consideration for structural
health monitoring systems. Ultrasonic signals transmitted and received between various array elements have been
shown to be effective for localizing discrete sites of damage using algorithms based upon changes in signals compared to
the undamaged state. Necessary to the success of the various imaging methods which have been proposed is a set of
baseline signals recorded under the same conditions as the signals acquired after damage has occurred. Since many
conditions other than structural damage can cause changes in ultrasonic signals, proposed here is an integrated strategy
whereby damage is first detected and is localized only if the outcome of the detection step is positive. In this manner
false alarms can be reduced since signal changes due to benign variations will not be localized and erroneously identified
as damage. The detection strategy considers the long time behavior of the signals in the diffuse-like regime where
distinct echoes can no longer be identified, whereas the localization strategy is to generate images of damage based upon
the early time regime when discrete echoes from boundary reflections and scattering sites are meaningful. Results are
shown for an aluminum plate subjected to a combination of temperature variations and introduction of artificial damage.
Ultrasonic methods are widely applied for nondestructive evaluation of structural components during both the
manufacturing process and subsequent field inspections. The field inspections often require expensive teardown in order
to access the back surfaces of critical components. Active ultrasonic methods are also a subject of ongoing research for
structural health monitoring whereby transducers are permanently attached to a structure and signals are monitored to
detect changes caused by structural damage. This paper presents a methodology for effectively combining ultrasonic
monitoring and inspection. During the monitoring phase, detection and localization of possible damage is demonstrated
on several specimens using attached transducers. This detection phase is followed by demonstration of a new inspection
method referred to as Acoustic Wavefield Imaging (AWI) which utilizes the attached transducers as sources and a single
externally scanned air-coupled transducer as a receiver. The acoustic wavefield images are useful for both checking the
viability of the attached transducers and quantifying the extent of damage. The AWI method approaches the sensitivity
of conventional through transmission ultrasonic methods but does not require access to both sides of structural
components. Thus, it is very well suited for rapid field inspection of structural assemblies.
Successful in-situ monitoring of crack initiation and growth is a necessary prerequisite for applying ultrasonic methods to structural health monitoring. For conventional ultrasonic testing methods, a focused beam may be used to directly image the crack tip; however, this method is difficult to apply during fatigue testing because of access limitations and couplant contamination issues. However, ultrasonic sensors can be permanently attached to a specimen to detect signal changes due to crack initiation and growth if the wave path is properly directed through the area of critical defect formation. The dynamics of cracks opening and closing during the fatigue process modulate the amplitude of ultrasonic waves propagating across these crack interfaces. Thus, even very small cracks can be reliably detected using permanently mounted sensors if the ultrasonic response can be measured as a function of load. A methodology is presented here that uses this behavior to detect and monitor crack formation and growth. This methodology may also be applied to structures subjected to unknown dynamic loads by using the ultrasonic signal to both estimate the instantaneous dynamic load and interrogate the integrity of the structure. Essential to the success of this method is an initial calibration on the undamaged structure where ultrasonic response is measured as a function of known static load. Results are presented from several aluminum specimens undergoing low cycle fatigue tests, and the dynamic loading results are shown to be comparable to the static ones in terms of the response of the ultrasonic signal to crack progression.
Changes in diffuse ultrasonic signals recorded from permanently mounted sensors can be correlated to initiation and growth of structural damage, offering hope that sparse sensor arrays can be utilized for monitoring large areas. It is well-known that benign environmental changes also have significant effects on diffuse ultrasonic signals that are of comparable magnitude to the effects of damage. Several methodologies are investigated for quantifying differences in diffuse ultrasonic signals by computing parameters that can be used to discriminate damage from environmental changes. The methodologies considered are waveform differencing, spectrogram differencing, change in local temporal coherence, and temperature compensated differencing. For all four methods, a set of baseline waveforms are first recorded from the undamaged specimen at a range of temperatures, and subsequently recorded waveforms are compared to those of the baseline set. Experimental data from aluminum plate specimens with artificial defects are analyzed. Results show that the local coherence method is the most effective for discriminating damage from temperature changes whereas waveform differencing is the least effective. Both the spectrogram differencing method and the temperature compensated differencing method offer intermediate performance. As expected, the efficacy of all four methods improves as the number of waveforms in the baseline set increases.
Diffuse ultrasonic signals received from ultrasonic sensors which are permanently mounted near, on or in critical structures of complex geometry are very difficult to interpret because of multiple modes and reflections constructively and destructively interfering. Both changing environmental and structural conditions affect the ultrasonic wave field, and the resulting changes in the received signals are similar and of the same magnitude. This paper describes a differential feature-based classifier approach to address the problem of determining if a structural change has actually occurred. Classifiers utilizing time and frequency domain features are compared to classifiers based upon time-frequency representations. Experimental data are shown from a metallic specimen subjected to both environmental changes and the introduction of artificial damage. Results show that both types of classifiers are successful in discriminating between environmental and structural changes. Furthermore, classifiers developed for one particular structure were successfully applied to a second one that was created by modifying the first structure. Best results were obtained using a classifier based upon features calculated from time-frequency regions of the spectrogram.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.