In ultrasonic structural health monitoring (SHM), environmental and operational conditions, especially temperature, can
significantly affect the propagation of ultrasonic waves and thus degrade damage detection. Typically, temperature
effects are compensated using optimal baseline selection (OBS) or optimal signal stretch (OSS). The OSS method
achieves compensation by adjusting phase shifts caused by temperature, but it does not fully compensate phase shifts
and it does not compensate for accompanying signal amplitude changes. In this paper, we develop a new temperature
compensation strategy to address both phase shifts and amplitude changes. In this strategy, OSS is first used to
compensate some of the phase shifts and to quantify the temperature effects by stretching factors. Based on stretching
factors, empirical adjusting factors for a damage indicator are then applied to compensate for the temperature induced
remaining phase shifts and amplitude changes. The empirical adjusting factors can be trained from baseline data with
temperature variations in the absence of incremental damage. We applied this temperature compensation approach to
detect volume loss in a thick wall aluminum tube with multiple damage levels and temperature variations. Our specimen
is a thick-walled short tube, with dimensions closely comparable to the outlet region of a frac iron elbow where flow-induced
erosion produces the volume loss that governs the service life of that component, and our experimental sequence
simulates the erosion process by removing material in small damage steps. Our results show that damage detection is
greatly improved when this new temperature compensation strategy, termed modified-OSS, is implemented.
“Clad steel” refers to a thick carbon steel structural plate bonded to a corrosion resistant alloy (CRA) plate, such as stainless steel or titanium, and is widely used in industry to construct pressure vessels. The CRA resists the chemically aggressive environment on the interior, but cannot prevent the development of corrosion losses and cracks that limit the continued safe operation of such vessels. At present there are no practical methods to detect such defects from the exposed outer surface of the thick carbon steel plate, often necessitating removing such vessels from service and inspecting them visually from the interior. In previous research, sponsored by industry to detect and localize damage in pressurized piping systems under operational and environmental changes, we investigated a number of data-driven signal processing methods to extract damage information from ultrasonic guided wave pitch-catch records. We now apply those methods to relatively large clad steel plate specimens. We study a sparse array of wafer-type ultrasonic transducers adhered to the carbon steel surface, attempting to localize mass scatterers grease-coupled to the stainless steel surface. We discuss conditions under which localization is achieved by relatively simple first-arrival methods, and other conditions for which data-driven methods are needed; we also discuss observations of plate-like mode properties implied by these results.
Sorting microparticles (or cells, or bacteria) is significant for scientific, medical and industrial purposes. Research groups have used lithium niobate SAW devices to produce standing waves, and then to align microparticles at the node lines in polydimethylsiloxane (PDMS, silicone) microfluidic channels. The “tilted angle” (skewed) configuration is a recent breakthrough producing particle trajectories that cross multiple node lines, making it practical to sort particles. However, lithium niobate wafers and PDMS microfluidic channels are not mechanically robust. We demonstrate “tilted angle” microparticle sorting in novel devices that are robust, rapidly prototyped, and manufacturable. We form our microfluidic system in a rigid polymethyl methacrylate (PMMA, acrylic) prism, sandwiched by lead-zirconium-titanate (PZT) wafers, operating in through-thickness mode with inertial backing, that produce standing bulk waves. The overall configuration is compact and mechanically robust, and actuating PZT wafers in through-thickness mode is highly efficient. Moving to this novel configuration introduced new acoustics questions involving internal reflections, but we show experimental images confirming the intended nodal geometry. Microparticles in “tilted angle” devices display undulating trajectories, where deviation from the straight path increases with particle diameter and with excitation voltage to create the mechanism by which particles are sorted. We show a simplified analytical model by which a “phase space” is constructed to characterize effective particle sorting, and we compare our experimental data to the predictions from that simplified model; precise correlation is not expected and is not observed, but the important physical trends from the model are paralleled in the measured particle trajectories.
The pulse-echo method is widely used for plate and pipe thickness measurement. However, the pulse echo method does not work well for detecting localized volumetric loss in thick-wall tubes, as created by erosion damage, when the morphology of volumetric loss is irregular and can reflect ultrasonic pulses away from the transducer, making it difficult to detect an echo. In this paper, we propose a novel method using an inductively coupled transducer to generate longitudinal waves propagating in a thick-wall aluminum tube for the volumetric loss quantification. In the experiment, longitudinal waves exhibit diffraction effects during the propagation which can be explained by the Huygens-Fresnel principle. The diffractive waves are also shown to be significantly delayed by the machined volumetric loss on the inside surface of the thick-wall aluminum tube. It is also shown that the inductively coupled transducers can generate and receive similar ultrasonic waves to those from wired transducers, and the inductively coupled transducers perform as well as the wired transducers in the volumetric loss quantification when other conditions are the same.
Guided waves can propagate long distances and are sensitive to subtle structural damage. Guided-wave based damage
localization often requires extracting the scatter signal(s) produced by damage, which is typically obtained by
subtracting an intact baseline record from a record to be tested. However, in practical applications, environmental and
operational conditions (EOC) dramatically affect guided wave signals. In this case, the baseline subtraction process can
no longer perfectly remove the baseline, thereby defeating localization algorithms.
In previous work, we showed that singular value decomposition (SVD) can be used to detect the presence of damage
under large EOC variations, because it can differentiate the trends of damage from other EOC variations. This capability
of differentiation implies that SVD can also robustly extract a scatter signal, originating from damage in the structure,
that is not affected by temperature variation. This process allows us to extract a scatterer signal without the challenges
associated with traditional temperature compensation and baseline subtraction routines. . In this work, we use to
approach to localize structural damage in large, spatially and temporally varying EOCs.
We collect pitch-catch records from randomly placed PZT transducers on an aluminum plate while undergoing
temperature variations. Damage is introduced to the plate during the monitoring period. We then use our SVD method
to extract the scatter signal from the records, and use the scatter signal to localize damage using the delay-and-sum
method. To compare results, we also apply several temperature compensation methods to the records and then perform
baseline subtraction. We show that our SVD-based approach successfully localize damage while current temperature-compensated
baseline subtraction methods fail.
Alkali-silica reaction (ASR) is a chemical reaction that can occur between alkaline components in cement paste and reactive forms of silica in susceptible aggregates when sufficient moisture is present. The ASR product, known as ASR gel, can cause expansion and cracking that damages the structure. We pass ultrasonic signals through concrete laboratory specimens and use three different ultrasonic methods to detect the onset of ASR damage, or the presence of ASR damage while still at the microscale. Our test specimens are fabricated with aggregate known to be reactive and are then exposed to an aggressive environment to accelerate ASR development. We use swept-sine excitations and obtain pitch-catch records from specimens that have been exposed to the accelerated environment. From this data, we demonstrate an ultrasonic passband method shows high frequency components diminish faster than low frequency components, and therefore the ultrasonic passband shifts to the low frequency side due to ASR damage. The test results also show that the ultrasonic passband is logically related to specimen size. We also demonstrate a stretching factor method is able to track the progress of ASR damage in concrete very well. These methods are shown to be more reliable than attenuation spectrum or attenuation methods that do not detect the ASR damage in concrete at early stages.
We describe lithium niobate SAW devices and their wave structure at different resonant frequencies, and we discuss the difference between PDMS and PMMA as the material for the microfluidic channel. We discuss the different wave structure for SAW devices operating at different resonant frequencies, showing simulation results and laboratory measurements. We discuss our recent studies to sort microparticles by size.
We describe lithium niobate SAW devices and PDMS microfluidic channels with which we study microparticle movement. We generate standing surface acoustic waves (with wavelengths of 200 micrometers) and show that microparticles (between 5 and 35 micrometers in diameter) move to nodes or antinodes. We report measurements of device response in the presence and absence of the microfluidic channel, which we combine with finite element simulation modeling to extract estimates of the PDMS damping.
Guided wave ultrasonics is an attractive technique for structural health monitoring, especially on pressurized pipes. However, civil infrastructure components, including pipes, are often subject to large environmental and operational variations that prevent traditional baseline subtraction-based approaches from detecting damage. We collect ultrasonic data on a large-scale pipe segment in its normal operating conditions and observe large environmental variations. We developed a damage detection method based on singular value decomposition (SVD) that is robust to those benign variations. We further develop an online novelty detection framework based on our SVD method to detect the presence of a mass scatterer on the pipe at the same time that we collect the data. We examine the framework with both synthetic simulations and field experimental data. The results show that the framework can effectively detect the presence of a scatterer and is robust to large environmental and operational variations.
We fabricated and tested (in a laboratory configuration) a flexible insert for wireless strain sensing in an intramedullary
tibial nail. The nail is a thick-walled titanium tube with an anatomic bend, and our flexible insert was designed to be
advanced into the hollow of the nail by a surgeon during the late stages of the fracture fixation process. The flexible
insert is a stainless steel rod, 170 mm long, with a part-circular section, roughly 4 mm wide and 2 mm deep that is
slightly smaller than the hollow. A lithium niobate SAW device developed by our research team, with an operating
frequency of 468 MHz, is bonded to the insert and demonstrates wireless strain sensing when the insert is bent.
We describe the mechanics of the flexible insert, which can be used for structural monitoring applications where
conventional strain gauge installation and wiring would be impractical. Because of its flexibility, the insert can be
advanced into an irregular channel, such as the hollow of a tibial nail, or a narrow access path drilled into a solid. The
insert will bend elastically into a configuration with at least three points of contact with the host body. The insert will be
prestrained (in bending) during installation and need not be anchored or bonded to the host body other than through
contact. Bending strain in the insert will vary as the host body deforms. We discuss possible application questions, and
we demonstrate strain sensing in a laboratory specimen.
Surface acoustic wave (SAW) devices are solid-state components in which a wave propagates along the surface of a
piezoelectric material. Changes in strain or temperature cause shifts in the acoustic wave speed and/or the path length,
enabling SAW devices to act as sensors. We present experimental studies on lithium niobate SAW devices acting as
passively-powered devices. Sensitivity, reproducibility, and linearity are excellent when measuring strain at constant
temperature, but the devices are also sensitive to temperature changes. We show experimental results of strain
measurement incorporating temperature compensation.
Strain monitoring is a nondestructive inspection method that can reveal the redistribution of internal forces, or the
presence of anomalous loadings, in structures. Surface acoustic wave (SAW) devices are small, robust, inexpensive
solid-state components in which a wave propagates along the surface of a piezoelectric material, and such devices are
used in large numbers commercially as delay devices and as filters. Changes in strain or temperature cause shifts in the
acoustic wave speed, by which such SAW devices can also serve as sensors. We present analytical, FEM simulation,
and experimental studies on SAW devices fabricated in our laboratory on lithium niobate wafers, with an inter-electrode
spacing of 8 micrometers. We discuss the change in wave speed with temperature and with strain, we outline the
influence of rotated cuts for the piezoelectric substrate, and we show results of laboratory sensing experiments.
Moreover, an electrode on a SAW device can be terminated as an antenna and interrogated with a wireless RF probe to
act as a passively-powered device, and we present laboratory results incorporating such wireless performance in our
research investigation. We pattern one set of electrodes on the SAW device as a transducer connected to the antenna,
and other sets of electrodes on the device acting as reflectors of the surface acoustic wave. At the RF frequencies used
for SAW devices, it is realistic to use directional antennas on the probe unit to achieve reasonable stand-off distances.
Our earlier research has studied the generation of nearly-longitudinal waves in thick plates by edge excitation at
relatively high frequency-thickness products. These nearly-longitudinal waves, also known as trailing pulses, are
promising for flaw detection due to their shorter wavelength and the capability of retaining the pulse characteristics after
scattering from defects. However, in reality, the edges of the structures may not be accessible. This paper explores
exciting ultrasonic waves using a wedge transducer at oblique incidence. We first describe a simple model for the
formation of trailing pulses by oblique excitation. We then use simulations to examine the creation of nearly-longitudinal
waves in a thick plate, study the effect of incidence angles and discuss the energy distribution through the thickness.
Next we provide experimental results from both pitch-catch and pulse-echo tests to validate the characteristics of the
nearly-longitudinal waves excited by oblique incidence. The good agreement between the simulation and experiments
shows oblique excitation can also produce nearly-longitudinal waves with uniformly-spaced trailing pulses over a range
of incidence angles. The amplitude level of such pulses reaches its maximum when the incidence angle approaches to the
critical angle at the plexiglass-steel interface. The responses by oblique excitation are weaker than those by edge
excitation, but can still illuminate the plate through the thickness when the incidence angle is close to the critical angle.
The results show that the nearly-longitudinal waves by oblique excitation are a good alternative for infrastructure
inspection, especially for plates limiting edge access or permitting only surface access.
Proc. SPIE. 7292, Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2009
KEYWORDS: Microelectromechanical systems, Signal to noise ratio, Electronics, Ferroelectric materials, Indium arsenide, Sensors, Amplifiers, Acoustic emission, Lead, Picture Archiving and Communication System
We present several findings related to acoustic emission detection using a MEMS sensor. The MEMS sensor is a
capacitive resonant transducer, fabricated in the PolyMUMPs process, with a resonant frequency near 160 kHz. In this
paper, the design, initial characterization, amplification electronics, and packaging of the sensor system are reviewed.
We present results from an experiment that compares the MEMS sensor system to a commercial PZT sensor by
comparing the response of each sensor to a pencil lead break on a plate. In addition, we describe the noise
characterization of the MEMS sensor system, comparing the predicted noise voltage to the measured value. This
analysis reveals that the electronic noise from the amplifier is significantly greater than the noise from the sensor,
suggesting that an amplifier with less noise would increase the sensitivity of the MEMS sensor system. We describe the
design of a new, transimpedance amplifier and its noise characterization, showing the new amplifier design has less
noise than the old design. The experiment comparing the commercial PZT sensor and the MEMS sensor system is
repeated using the new amplifier, and we present results showing an increase in sensitivity of the MEMS sensor system.
Finally, we present results from an experiment comparing the ability of the commercial PZT sensor and the MEMS
sensor system with the new amplifier to detect pencil lead breaks performed a large distance from each sensor.
A steel box beam in a monorail application is constructed with an epoxy grout wearing surface, precluding visual
inspection of its top flange. This paper describes a sequence of experimental research tasks to develop an ultrasonic
system to detect flaws (such as fatigue cracks) in that flange, and the results of a field test to demonstrate system
performance. The problem is constrained by the fact that the flange is exposed only along its longitudinal edges, and by
the fact that permanent installation of transducers at close spacing was deemed to be impractical. The system chosen for
development, after experimental comparison of alternate technologies, features angle-beam ultrasonic transducers with
fluid coupling to the flange edge; the emitting transducers create transverse waves that travel diagonally across the width
of the flange, where an array of receiving transducers detect flaw reflections and flaw shadows. The system rolls along
the box beam, surveying (screening) the top flange for the presence of flaws.
In a first research task, conducted on a full-size beam specimen, we compared waves generated from different transducer
locations, either the flange edge or the web face, and at different frequency ranges. At relatively low frequencies, such
as 100 kHz, we observed Lamb wave modes, and at higher frequency, in the MHz range, we observed nearlylongitudinal
waves with trailing pulses. In all cases we observed little attenuation by the wearing surface and little
influence of reflection at the web-flange joints. At the conclusion of this task we made the design decision to use edgemounted
transducers at relatively high frequency, with correspondingly short wavelength, for best scattering from flaws.
In a second research task we conducted experiments at 55% scale on a steel plate, with machined flaws of different size,
and detected flaws of target size for the intended application. We then compared the performance of bonded transducers,
fluid-coupled transducers, and angle-beam (wedge) transducers; from that comparison we made the design decision to
use wedges, which beam the wave to increase the scattering from flaws. We also compared the performance of wired
transducers using fluid coupling to that of wireless (inductively coupled) transducers mounted permanently. Although
the wireless transducers achieved flaw detection, the necessary spacing (determined experimentally) would have
required an impractical number of transducers. Therefore, we made the design decision to use wedge transducers with
In a third research task we developed and tested a rolling system with a water channel for acoustic coupling, including a
study of its sensitivity to misalignment, and in a fourth task we devised a data display to create a pattern of reflections or
shadows that could be easily interpreted as evidence of a flaw. Finally, we conducted a field test on the full-size system
in a region containing bolt holes, which act as a physical simulation of a flaw, and show successful detection of
reflections and shadows from those holes.
We present four new findings pertaining to MEMS sensors for acoustic emission detection. Our sensors are resonant-type
capacitive transducers, operating with a frequency between 100 kHz and 500 kHz, fabricated in the PolyMUMPS
process. The sensitivity of a resonant transducer is related to the sharpness of its resonance, measured by the quality
factor Q, and operating in a coarse vacuum will increase Q. We describe a practical laboratory method for sealing and
evacuating our MEMS sensor, and present measurements showing Q in the evacuated packages to be 2.4 to 3.6 times
greater than under atmospheric pressure. We also describe our theoretical analysis of noise sources in the
electromechanical behavior of a resonant, capacitive-type transducer sensitive to out-of-plane motion, with particular
interest in noise resulting from mechanical excitation of the moving plate by air molecule impact. We report on a new
transducer design to sense out-of-plane motion featuring a moving plate constructed as an open grill rather than as a
plate perforated by etch holes. Characterization measurements show the open grill design to have a higher Q than a
comparable perforated plate transducer. Finally, we report on another novel transducer designed to sense in-plane
motion. The sensor is a comb finger capacitive transducer, and theoretical predications predict the in-plane sensor to
have a much higher Q than the out-of-plane sensors. We show experimental measurements confirming these design
characteristics, and we show results from pencil lead break experiments.
We discuss waves created in relatively thick plates by edge excitation at frequency-thickness (fd) products that
correspond, in principle, to multiple Lamb wave modes. For relatively low values of the fd product it is clear that Lamb
wave modes will be generated, while at large values of the fd product we observe a bulk (longitudinal wave) in the solid,
but influenced by reflections from the plate surfaces. We show that for a range of intermediate fd products a train of
regularly spaced nearly-longitudinal waves is generated. The development of a lead pulse and trailing pulses, all
traveling at the longitudinal (bulk) wave speed, is well known and has been explained in the literature. In this paper we
describe the transition from Lamb wave generation to the formation of nearly-longitudinal waves with their trailing
pulses. We report experimental results and theoretical results, with good correspondence between them. We also
examine the transfer of energy from leading to trailing pulses, which means that such nearly-longitudinal waves will not
propagate indefinitely; however, we show that they retain ample energy for flaw detection at distances of several meters.
Most importantly, we study the interaction of these trailing pulses with cracks, again showing experimental results and
theoretical predictions that are consistent with one another. The results suggest that these nearly-longitudinal waves are
an attractive option for flaw detection because of their shorter wavelength (as compared to Lamb waves at low fd
products) and because they preserve their pulse train characteristics after scattering.
In earlier work we developed inductive coupling for surface-mounted Lamb wave transducers operating at relatively low
frequencies, such as 300 kHz. We now report on similar inductively-coupled transducers at higher (multiple MHz)
frequencies. Our investigation was motivated by a particular application, to examine a box girder top flange, using
transducers mounted along the flange edge. We employ a pair of transducers in pitch-catch mode, offset to create a
diagonal path, and show that a shadow is detected when the path is intercepted by a through-thickness crack. We
compare results obtained using conventionally wired transducers and using inductively-coupled transducers, showing
that effective performance can be achieved with wireless (inductively-coupled) operation. Superior performance is
obtained if plexiglas wedges are used to direct the beam along the diagonal path. Reflection from the crack is evident, as
is the shadow effect along the direct diagonal path.
A novel transducer for active or passive sensing has been developed and tested experimentally. It features a steel wire
acting as a wave guide between a piezoceramic element and the structure under test. Some advantages of the wire-guided
transducer include its applicability to structures operating at high temperature, which otherwise preclude the
surface mounting of piezoceramics, its small contact area to the structure, which enables several such transducers to be
deployed in an arc around a known crack location as an acoustic emission sensor array, and its low cost and ease of
installation. Another potential advantage is simplified signal processing for source localization, which is developed in
this paper and evaluated experimentally.
The various steel wires used in our experiments to date are less than 1 mm in diameter and between 10 cm and 100 cm in
length. The wire guides have been studied with active excitation under a pulse excitation as used in ultrasonic testing, at
a relatively high frequency such as 1 MHz, and in the frequency range of 100 kHz to 500 kHz which is often of interest
for Lamb wave generation in thin plates or for acoustic emission sensing. Our tests confirm that the wire acts as a
cylindrical rod in which the fastest wave is the lowest longitudinal mode, displaying a sharp arrival, and in which the
lowest flexural mode and lowest torsional mode are also excited; we report excellent agreement between measured and
predicted wave speeds, as expected.
We show experimental results in which a group of wire-guided transducers permit the localization of an impact on a thin
plate and discuss the automation of this task for use in the field. We also show the ability of the wire-guided transducer
to detect acoustic emission events simulated physically by pencil lead breaks.
We report on the application of wafer-type PZT transducers to the detection of flaws in steel plate girders. In these
experiments one transducer is used to emit a pulse and the second receives the pulse and reflections from nearby
boundaries, flaws, or discontinuities (pitch-catch mode). In this application there will typically be numerous reflections
observed in the undamaged structure. A major challenge is to recognize new reflections caused by fatigue cracks in the
presence of these background reflections. A laboratory specimen plate girder was fabricated at approximately half scale,
910 mm deep with an h/t ratio of 280 for the web and a b/t ratio of 16 for the flanges, and with transverse stiffeners
fabricated with a web gap at the tension flange. Two wafer-type transducers were mounted on the web approximately
175 mm from the crack location, one on each side of the stiffener. The transducers were operated in pitch-catch mode,
excited by a windowed sinusoid to create a narrowband transient excitation. The transducer location relative to the crack
corresponded to a total included angle of roughly 30 degrees in the path reflecting from the crack. Cyclic loading was
applied to develop a distortion-induced fatigue crack in the web at the web gap location. After appearance of the crack,
ultrasonic measurements were performed at a range of center frequencies below the cutoff frequency of the A1 Lamb
wave mode. Subsequently the crack was extended mechanically to simulate crack growth under primary longitudinal
(bending) stress and the measurements were repeated. Direct differencing of the signals showed arrivals at times
corresponding to reflection from the crack location, growing in amplitude as the crack was lengthened mechanically.
These results demonstrate the utility of Lamb waves for crack detection even in the presence of numerous background
An improved multi-channel MEMS chip for acoustic emission sensing has been designed and fabricated in 2006 to
create a device that is smaller in size, superior in sensitivity, and more practical to manufacture than earlier designs. The
device, fabricated in the MUMPS process, contains four resonant-type capacitive transducers in the frequency range
between 100 kHz and 500 kHz on a chip with an area smaller than 2.5 sq. mm. The completed device, with its circuit
board, electronics, housing, and connectors, possesses a square footprint measuring 25 mm x 25 mm. The small
footprint is an important attribute for an acoustic emission sensor, because multiple sensors must typically be arrayed
around a crack location. Superior sensitivity was achieved by a combination of four factors: the reduction of squeeze
film damping, a resonant frequency approximating a rigid body mode rather than a bending mode, a ceramic package
providing direct acoustic coupling to the structural medium, and high-gain amplifiers implemented on a small circuit
board. Manufacture of the system is more practical because of higher yield (lower unit costs) in the MUMPS fabrication
task and because of a printed circuit board matching the pin array of the MEMS chip ceramic package for easy assembly
The transducers on the MEMS chip incorporate two major mechanical improvements, one involving squeeze film
damping and one involving the separation of resonance modes. For equal proportions of hole area to plate area, a
triangular layout of etch holes reduces squeeze film damping as compared to the conventional square layout. The effect
is modeled analytically, and is verified experimentally by characterization experiments on the new transducers.
Structurally, the transducers are plates with spring supports; a rigid plate would be the most sensitive transducer, and
bending decreases the sensitivity. In this chip, the structure was designed for an order-of-magnitude separation between
the first and the second mode frequency, strongly approximating the desirable rigid plate limit. The effect is modeled
analytically and is verified experimentally by measurement of the resonance frequencies in the new transducers.
Another improvement arises from the use of a pin grid array ceramic package, in which the MEMS chip is acoustically
coupled to the structure with only two interfaces, through a ceramic medium that is negligible in thickness when
compared to wavelengths of interest.
Like other acoustic emission sensors, those on the 2006 MEMS chip are sensitive only to displacements normal to the
surface on which the device is mounted. To overcome that long-standing limitation, a new MEMS sensor sensitive to in-plane
motion has been designed, featuring a different spring-mass mechanism and creating the signal by the change in
capacitance between stationary and moving fingers. Predicted damping is much lower for the case of the in-plane
sensor, and squeeze-film damping is used selectively to isolate the desired in-plane mechanical response from any
unwanted out-of-plane response. The new spring-mass mechanism satisfies the design rules for the PolyMUMPS
fabrication (foundry) process. A 3-D MEMS sensor system is presently being fabricated, collocating two in-plane
sensors and one out-of-plane sensor at the mm scale, which is very short compared to the acoustic wavelength of interest
for stress waves created by acoustic emission events.
Lamb waves in plates and in cylindrical pipes have been the subject of extensive study, largely because they propagate great distances with little attenuation, and can therefore be used to detect flaws. In this paper we report finite element simulations and experimental studies of Lamb waves in steel bridge girder geometries. In our studies the Lamb waves are generated by PZT wafer-type transducers mounted on the girder web, driven by a windowed sinusoidal pulse; the pulse center frequency is chosen to yield a frequency-thickness product of roughly 1 MHz-mm, at which the group velocities of the S0 and A0 waves are well separated, and at which waves in higher modes are theoretically absent. Transient dynamic finite element simulations, both in 2D and in 3D, were performed using FEMLAB and ABAQUS. The simulations show that transmission at the web-flange joint creates guided waves in the flanges that travel at different velocities from the Lamb waves in the web, and that reflection at the web-flange joint creates a largely straight-crested wavefront for the Lamb waves in the web remote from the source. Simulation studies also illustrate the acoustic influence of plate girder transverse stiffeners, which is observed to be relatively small. A welded steel plate girder laboratory specimen was fabricated with proportions typical of highway bridge members, at approximately half-scale. The web height is 920 mm and thickness is 3.2 mm, for a representative height-thickness ratio of 288; the flange width is 100 mm and thickness is 6.4 mm, for a representative width-thickness ratio of 16. Small PZT transducers, roughly 6.4 x 6.4 x 0.6 mm, excited at less than 10 V, produce ample signals. We compare simulation results and experimental measurements for Lamb wave illumination of the plate girder segment. We also discuss the detection of cracks, simulated experimentally by saw cuts of varying dimensions in the laboratory girder specimen.
In a collaborative project at Lehigh and Carnegie Mellon, a MEMS acoustic emission sensor was designed and fabricated as a suite of six resonant-type capacitive transducers in the frequency range between 100 and 500 kHz. Characterization studies showed good comparisons between predicted and experimental electro-mechanical behavior. Acoustic emission events, simulated experimentally in steel ball impact and in pencil lead break tests, were detected and source localization was demonstrated. In this paper we describe the application of the MEMS device in structural testing, both in laboratory and in field applications. We discuss our findings regarding housing and mounting (acoustic coupling) of the MEMS device with its supporting electronics, and we then report the results of structural testing.
In all tests, the MEMS transducers were used in parallel with commercial acoustic emission sensors, which thereby serve as a benchmark and permit a direct observation of MEMS device functionality. All tests involved steel structures, with particular interest in propagation of existing cracks or flaws. A series of four laboratory tests were performed on beam specimens fabricated from two segments (Grade 50 steel) with a full penetration weld (E70T-4 electrode material) at midspan. That weld region was notched, an initial fatigue crack was induced, and the specimens were then instrumented with one commercial transducer and with one MEMS device; data was recorded from five individual transducers on the MEMS device. Under a four-point bending test, the beam displayed both inelastic behavior and crack propagation, including load drops associated with crack instability. The MEMS transducers detected all instability events as well as many or most of the acoustic emissions occurring during plasticity and stable crack growth. The MEMS transducers were less sensitive than the commercial transducer, and did not detect as many events, but the normalized cumulative burst count obtained from the MEMS transducers paralleled the count obtained from the commercial transducer. Waveform analysis of signals from the MEMS transducers provided additional information concerning arrivals of P-waves and S-waves. Similarly, the analysis provided additional confirmation that the acoustic emissions emanated from the damage zone near the crack tip, and were not spurious signals or artifacts.
Subsequent tests were conducted in a field application where the MEMS transducers were redundant to a group of commercial transducers. The application example is a connection plate in truss bridge construction under passage of heavy traffic loads. The MEMS transducers were found to be functional, but were less sensitive in their present form than existing commercial transducers. We conclude that the transducers are usable in their current configuration and we outline applications for which they are presently suited, and then we discuss alternate MEMS structures that would provide greater sensitivity.
Lamb waves at ultrasonic frequencies travel with little attenuation in thin elastic plates, and we demonstrate their use in pulse-echo behavior to monitor plate integrity. We envision using a single PZT wafer-type transducer to generate waves and to receive reflections from distant flaw or boundary locations. However, Lamb waves generally have multiple modes, each of them highly dispersive, and in consequence pulse dispersion can become pronounced and can make difficult or impossible the interpretation of pulse-echo responses. We show that selective generation of the S0 wave will overcome those difficulties; therefore, selection of transducer dimensions and pulse characteristics to achieve selective generation should be considered mandatory for most intended applications. We first review the work of others identifying a basic relationship between transducer dimension and excitation frequency for selective generation of the S0 wave. We then summarize our extensive experimental studies of wafer-type transducers with particular attention to S0 and A0 mode behavior, both in transmission and reception. We next report our two-dimensional finite element simulation of the same problem performed in FEMLAB, requiring transient simulation of the coupled electromechanical problem. We simulate the piezoelectric response of the wafer-type transducer coupled to the elastic plate, both as transmitter and receiver, as well as the development of Lamb waves within the source region and their subsequent propagation along the plate. Simulations illustrate the development and separation of the S0 and A0 modes and reproduce the expected group velocities and dispersion behavior. We show good agreement between our experiments and our simulations regarding S0 mode behavior, and we summarize the results to guide a designer in choosing transducer dimensions. In particular, good selectivity between the S0 and A0 mode generation can be obtained with appropriate choice of transducer size and center frequency. We show the results of experiments on an aluminum plate in which excitation of a single PZT wafer-type transducer at 6.5 V (peak-to-peak) produces reflected signals of ample strength (tens of mV) from distant boundaries and from partial thickness flaws.
Proc. SPIE. 5765, Smart Structures and Materials 2005: Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems
KEYWORDS: Microelectromechanical systems, Etching, Capacitance, Transducers, Acoustic emission, Signal detection, Atmospheric sensing, Surface micromachining, Lead, Picture Archiving and Communication System
We describe the design, fabrication, testing and application (in structural experiments) of our 2004 (second generation) MEMS device, designed for acoustic emission sensing based upon experiments with our 2002 (first generation) device. Both devices feature a suite of resonant-type transducers in the frequency range between 100 kHz and 1 MHz. The 2002 device was designed to operate in an evacuated housing because of high squeeze film damping, as confirmed in our earlier experiments. In additional studies involving the 2002 device, experimental simulation of acoustic emissions in a steel plate, using pencil lead break or ball impact loading, showed that the transducers in the frequency range of 100 kHz-500 kHz presented clearer output signals than the transducers with frequencies higher than 500 kHz. Using the knowledge gained from the 2002 device, we designed and fabricated our second generation device in 2004 using the multi-user polysilicon surface micromachining (MUMPs) process. The 2004 device has 7 independent capacitive type transducers, compared to 18 independent transducers in the 2002 device, including 6 piston type transducers in the frequency range of 100 kHz to 500 kHz and 1 piston type transducer at 1 MHz to capture high frequency information. Piston type transducers developed in our research have two uncoupled modes so that twofold information can be acquired from a single transducer. In addition, the piston shape helps to reduce residual stress effect of surface micromachining process. The center to center distance between etch holes in the vibrating plate was reduced from 30 μm to 13 μm, in order to reduce squeeze film damping. As a result, the Q factor under atmospheric pressure for the 100 kHz transducer was increased to 2.37 from 0.18, and therefore the vacuum housing has been eliminated from the 2004 device. Sensitivities of transducers were also increased, by enlarging transducer area, in order to capture significant small amplitude acoustic emission events. The average individual transducer area in the 2004 device was increased to 6.97 mm2 as compared to 2.51 mm2 in the 2002 device. In this paper, we report the new experimental results on the characterization of the 2004 device and compare them with analytical results. We show improvements in sensitivity as measured by capacitance and as measured by pencil lead break experiments. Improvement in damping is also evaluated by admittance measurement in atmosphere. Pencil lead break experiments also show that transducers can operate in atmospheric pressure. Finally, we apply the device to acoustic emission experiments on crack propagation in a steel beam specimen, precracked in fatigue, in a four-point bending test.
MEMS ultrasonic transducers for flaw detection have heretofore been built as capacitive diaphragm-type devices. A diaphragm forms a moveable electrode, placed at a short gap from a stationary electrode, and diaphragm movement has been detected by capacitance change. Although several research teams have successfully demonstrated that technology, the detection of capacitance change is adversely affected by stray and parasitic capacitances, limiting the sensitivity of such transducers and typically requiring relatively large diaphragm areas. We describe the design and fabrication of what to our knowledge is the first CMOS-MEMS ultrasonic phased array transducer using piezoresistive strain sensing. Piezoresistors have been patterned within the diaphragms, and diaphragm movement creates bending strain which is detected by a bridge circuit, for which conductor losses will be less significant. The prospective advantage of such piezoresistive transducers is that sufficient sensitivity may be achieved with very small diaphragms. We compare transducer response under fluid-coupled ultrasonic excitation and report the experimental gauge factor for the piezoresistors. We also discuss the phased array performance of the transducer in sensing the direction of an incoming wave.
Acoustic emission testing is a passive nondestructive testing technique used to identify the onset and characteristics of damage through the detection and analysis of transient stress waves. Successful detection and implementation of acoustic emission requires good coupling, high transducer sensitivity and ability to discriminate noise from real signals. We report here detection of simulated acoustic emission signals using a MEMS chip fabricated in the multi-user polysilicon surface micromachining (MUMPs) process. The chip includes 18 different transducers with 10 different resonant frequencies in the range of 100 kHz to 1 MHz. It was excited by two different source simulation techniques; pencil lead break and impact loading. The former simulation was accomplished by breaking 0.5 mm lead on the ceramic package. Four transducer outputs were collected simultaneously using a multi-channel oscilloscope. The impact loading was repeated for five different diameter ball bearings. Traditional acoustic emission waveform analysis methods were applied to both data sets to illustrate the identification of different source mechanisms. In addition, a sliding window Fourier transform was performed to differentiate frequencies in time-frequency-amplitude domain. The arrival and energy contents of each resonant frequency were investigated in time-magnitude plots. The advantages of the simultaneous excitation of resonant transducers on one chip are discussed and compared with broadband acoustic emission transducers.
Acoustic emission testing is an important technology for evaluating structural materials, and especially for detecting damage in structural members. Significant new capabilities may be gained by developing MEMS transducers for acoustic emission testing, including permanent bonding or embedment for superior coupling, greater density of transducer placement, and a bundle of transducers on each device tuned to different frequencies. Additional advantages include capabilities for maintenance of signal histories and coordination between multiple transducers. We designed a MEMS device for acoustic emission testing that features two different mechanical types, a hexagonal plate design and a spring-mass design, with multiple detectors of each type at ten different frequencies in the range of 100 kHz to 1 MHz. The devices were fabricated in the multi-user polysilicon surface micromachining (MUMPs) process and we have conducted electrical characterization experiments and initial experiments on acoustic emission detection. We first report on C(V) measurements and perform a comparison between predicted (design) and measured response. We next report on admittance measurements conducted at pressures varying from vacuum to atmospheric, identifying the resonant frequencies and again providing a comparison with predicted performance. We then describe initial calibration experiments that compare the performance of the detectors to other acoustic emission transducers, and we discuss the overall performance of the device as a sensor suite, as contrasted to the single-channel performance of most commercial transducers.
In earlier work we developed a MEMS phased array transducer, fabricated in the MUMPs process, and we reported on initial experimental studies in which the device was affixed into contact with solids. We demonstrated the successful detection of signals from a conventional ultrasonic source, and the successful localization of the source in an off-axis geometry using phased array signal processing. We now describe the predicted transmission and coupling characteristics for such devices in contact with solids, demonstrating reasonable agreement with experimental behavior. We then describe the results of flaw detection experiments, as well as results for fluid-coupled detectors.
Resident sensors are envisioned for civil infrastructure applications, but providing long-term power for such devices remains a design challenge. An ideal solution would be to scavenge energy from structural strains and make the scavenging component a part of the sensor package. In principle, piezoelectric materials are suited to that role, and studies by others have demonstrated the feasibility of energy scavenging from flexible PZT devices operated at large strains and high strain rates. We have conducted experiments to collect electrical energy from PZT ceramics. We summarize the governing piezoelectric equations and outline the most convenient forms to use for the energy scavenging problem, illustrated by tracing one complete loading cycle. We review the material properties for the three PZT ceramics used in our experiments. We show experimental results recording voltage and charge in the cases of open-circuit, resistive loads, and capacitive loads, showing good agreement with analytical predictions. However, the greatest challenge is the approach to energy storage. In theory, capacitors can store energy but at varying voltage and with non-negligible leakage, whereas a battery can store energy at constant voltage with little leakage. We conducted experiments on both approaches, and we discuss our findings of the feasibility and efficiency of battery recharging at the scale of our devices, which have nominal dimensions of 10x10x1 mm.
Ultrasonic methods can be used to monitor crack propagation, weld failure, or section loss at critical locations in steel structures. However, ultrasonic inspection requires a skilled technician, and most commonly the signal obtained at any inspection is not preserved for later use. A preferred technology would use a MEMS device permanently installed at a critical location, polled remotely, and capable of on-chip signal processing using a signal history. We review questions related to wave geometry, signal levels, flaw localization, and electromechanical design issues for microscale transducers, and then describe the design, characterization, and initial testing of a MEMS transducer to function as a detector array. The device is approximately 1-cm square and was fabricated by the MUMPS process. The chip has 23 sensor elements to function in a phased array geometry, each element containing 180 hexagonal polysilicon diaphragms with a typical leg length of 49 microns and an unloaded natural frequency near 3.5 MHz. We first report characterization studies including capacitance-voltage measurements and admittance measurements, and then report initial experiments using a conventional piezoelectric transducer for excitation, with successful detection of signals in an on-axis transmission experiment and successful source localization from phased array performance in an off-axis transmission experiment.