An integral aspect of modern infrastructural engineering is to constantly monitor the health of a structure either actively or passively in order to ensure its safe performance throughout the design life. For passive structural health monitoring, it is important to estimate the location of an acoustic source that may be caused by events such as impact of a foreign object with the structure, failure of a structural element, formation of cracks, etc. Such an acoustic source generates acoustic waves that propagate through the medium. These waves can be captured by ultrasonic sensors mounted on the structure at some pre-selected locations and, subsequently, analyzed to predict the location of the acoustic source. Over the years, several researchers have proposed techniques for acoustic source localization in both isotropic and anisotropic structures. While acoustic source localization in isotropic structures is relatively simple, introduction of anisotropy adds a layer of difficulty to the problem due to the fact that waves do not propagate with the same speed in all directions. This study presents acoustic source localization techniques for anisotropic plates based on the analysis of the wave front shapes typically observed in anisotropic plates and presents experimental verification of the techniques. Three different geometric shapes are considered as the assumed wave front shapes: a rhombus, an ellipse and a parametric curve. A slightly modified version of the rhombus-based technique from the original approach is proposed. The experimental study is performed on two plates with different degrees of anisotropy.
Carbon nanotube-based composites have been deeply investigated in recent years. CNTs due to their unique physical properties have been employed for various applications in different disciplines of science and engineering. Due to the remarkable electrical conductivity of CNTs, one of practical applications is related to the development of strain sensing smart coatings. Up to now, high sensitivity strain sensors for micro- and macro-scale applications were proposed. However, controlling electro-mechanical properties of those devices is still a challenging issue.
In order to facilitate the design process and to investigate multi-domain relationships between sensor's parameters and its properties, numerical models and simulations of CNT-based structures have been carried out with the primary focus on investigation of electrical conductivity for various concentrations of CNTs within the composite material. More accurate and detailed studies include analysis of the influence of deformation on changes in conductivity. However, due to significant microstructural complexity of the system (i.e. large number of CNTs within the structure) multiscale modeling and analysis approach must be employed.
The main objective of this paper is to outline an relationships between micro- and macro-structural properties of CNT-based strain sensors and to discuss guidelines for a multiscale electro-mechanical model based on the Representative Volume Element (RVE) concept. The device employs the change in electrical conductivity of a CNT-based nanocomposite under applied deformation. The study investigates the impact of various micro-scale model parameters (e.g. size of an RVE, CNTs parameters etc.), on the macro scale model. Model parameters convergence studies are performed for different geometrical properties of CNTs and for various sizes of RVEs - revealing their critical mutual relationships. Also, the impact of boundary conditions at the micro-scale RVE structure is discussed.
Nonlinear features of elastic wave propagation have attracted significant attention recently. The particular interest herein relates to complex wave-structure interactions, which provide potential new opportunities for feature discovery and identification in a variety of applications. Due to significant complexity associated with wave propagation in nonlinear media, numerical modeling and simulations are employed to facilitate design and development of new measurement, monitoring and characterization systems. However, since very high spatio- temporal accuracy of numerical models is required, it is critical to evaluate their spectral properties and tune discretization parameters for compromise between accuracy and calculation time. Moreover, nonlinearities in structures give rise to various effects that are not present in linear systems, e.g. wave-wave interactions, higher harmonics generation, synchronism and | recently reported | shifts to dispersion characteristics. This paper discusses local computational model based on a new HYBRID approach for wave propagation in nonlinear media. The proposed approach combines advantages of the Local Interaction Simulation Approach (LISA) and Cellular Automata for Elastodynamics (CAFE). The methods are investigated in the context of their accuracy for predicting nonlinear wavefields, in particular shifts to dispersion characteristics for finite amplitude waves and secondary wavefields. The results are validated against Finite Element (FE) calculations for guided waves in copper plate. Critical modes i.e., modes determining accuracy of a model at given excitation frequency - are identified and guidelines for numerical model parameters are proposed.
Nonlinear ultrasonic techniques provide improved damage sensitivity compared to linear approaches. The combination of attractive properties of guided waves, such as Lamb waves, with unique features of higher harmonic generation provides great potential for characterization of incipient damage, particularly in plate-like structures. Nonlinear ultrasonic structural health monitoring techniques use interrogation signals at frequencies other than the excitation frequency to detect changes in structural integrity. Signal processing techniques used in non-destructive evaluation are frequently supported by modeling and numerical simulations in order to facilitate problem solution. This paper discusses known and newly-developed local computational strategies for simulating elastic waves, and attempts characterization of their numerical properties in the context of linear and nonlinear media. A hybrid numerical approach combining advantages of the Local Interaction Simulation Approach (LISA) and Cellular Automata for Elastodynamics (CAFE) is proposed for unique treatment of arbitrary strain-stress relations. The iteration equations of the method are derived directly from physical principles employing stress and displacement continuity, leading to an accurate description of the propagation in arbitrarily complex media. Numerical analysis of guided wave propagation, based on the newly developed hybrid approach, is presented and discussed in the paper for linear and nonlinear media. Comparisons to Finite Elements (FE) are also discussed.
The application of guided waves using surface-bonded piezoceramic transducers for nondestructive testing (NDT) and Structural Health Monitoring (SHM) have shown great potential. However, due to difficulty in identification of individual wave modes resulting from their dispersive and multi-modal nature, selective mode excitement methods are highly desired.
The presented work focuses on an optimization-based approach to design of a piezoelectric transducer for selective guided waves generation. The concept of the presented framework involves a Finite Element Method (FEM) model in the optimization process. The material of the transducer is optimized in topological sense with the aim of tuning piezoelectric properties for actuation of specific guided wave modes.
Remarkable electrical properties of carbon nanotubes (CNT) have lead to increased interest in studying CNT- based devices. Many of current researches are devoted to using all kinds of carbon nanomaterials in the con- struction of sensory elements. One of the most common applications is the development of high performance, large scale sensors. Due to the remarkable conductivity of CNT's such devices represent very high sensitivity. However, there are no sufficient tools for studying and designing such sensors. The main objective of this paper is to develop and validate a multiscale numerical model for a carbon nanotubes based sensor. The device utilises the change of electrical conductivity of a nanocomposite material under applied deformation. The nanocomposite consists of a number of CNTs dispersed in polymer matrix. The paper is devoted to the analysis of the impact of spatial distribution of carbon nanotubes in polymer matrix on electrical conductivity of the sensor. One of key elements is also to examine the impact of strain on electric charge ow in such anisotropic composite structures. In the following work a multiscale electro-mechanical model for CNT - based nanocomposites is proposed. The model comprises of two length scales, namely the meso- and the macro-scale for mechanical and electrical domains. The approach allows for evaluation of macro-scale mechanical response of a strain sensor. Electrical properties of polymeric material with certain CNT fractions were derived considering electrical properties of CNTs, their contact and the tunnelling effect.
A number of techniques are available for acoustic source localization in isotropic plates without knowing the material properties of the plate. However, for a highly anisotropic plate acoustic source localization requires some knowledge of the plate material properties or its group velocity profile. In absence of this information one requires a large number of sensors to predict the acoustic source point in the plate. All proposed techniques for acoustic source localization with a few sensors assume the straight line propagation of waves from the source to the receiving sensor with an average group velocity when the plate material properties are not known. However, this assumption is not true for an anisotropic plate. Although the currently available techniques work well for weakly anisotropic plates since the wave path does not deviate significantly from the straight line propagation they fail miserably for highly anisotropic plates.
In this paper acoustic source is localized in an anisotropic plate when non-circular wave front is generated. Direction vectors of wave fronts are obtained from the Time-Difference-Of-Arrivals (TDOA) at three sensors placed in a cluster. Four such direction vectors are then utilized in geometric vector analysis to accurately obtain the acoustic source location. The proposed technique is illustrated on an orthotropic plate that generates rhombus shaped wave front. It should be noted that the proposed technique does not require wave propagation along a straight
Acoustic emission is a vital non-destructive testing technique and is widely used in industry for damage detection, localisation and characterization. The latter two aspects are particularly challenging, as AE data are typically noisy. What is more, elastic waves generated by an AE event, propagate through a structural path and are significantly distorted. This effect is particularly prominent for thin elastic plates. In these media the dispersion phenomenon results in severe localisation and characterization issues. Traditional Time Difference of Arrival methods for localisation techniques typically fail when signals are highly dispersive. Hence, algorithms capable of dispersion compensation are sought. This paper presents a method based on the Time - Distance Domain Transform for an accurate AE event localisation. The source localisation is found through a minimization problem. The proposed technique focuses on transforming the time signal to the distance domain response, which would be recorded at the source. Only, basic elastic material properties and plate thickness are used in the approach, avoiding arbitrary parameters tuning.
Nonlinear elastic guided waves find application in various disciplines of science and engineering, such as non- destructive testing and structural health monitoring. Recent recognition and quantification of their amplitude- dependent changes in spectral properties has contributed to the development of new monitoring concepts for mechanical structures. The focus of this work is to investigate and predict amplitude-dependent shifts in Lamb wave dispersion curves. The theory for frequency/wavenumber shifts for plate waves, based on a Lindstedt-Poincaré perturbation approach, was presented by the authors in previous years. Equivalently, spectral properties changes can be seen as wavelength contraction/elongation. Within the proposed framework, the wavelength of a Lamb wave depends on several factors; e.g., wave amplitude and second-, third- and fourth-order elastic constants, and others. Various types of nonlinear effects are considered in presented studies. Sensitivity studies for model parameters, i.e. higher-order elastic constants, are performed to quantify their influence on Lamb wave frequency/wavenumber shifting, and to identify the key parameters governing wavelength tuning.
Attractive properties of guided waves provides very unique potential for characterization of incipient damage, particularly in plate-like structures. Among other properties, guided waves can propagate over long distances and can be used to monitor hidden structural features and components. On the other hand, guided propagation brings substantial challenges for data analysis. Signal processing techniques are frequently supported by numerical simulations in order to facilitate problem solution. When employing numerical models additional sources of errors are introduced. These can play significant role for design and development of a wave-based monitoring strategy. Hence, the paper presents an investigation of numerical models for guided waves generation, propagation and sensing. Numerical dispersion analysis, for guided waves in plates, based on the LISA approach is presented and discussed in the paper. Both dispersion and modal amplitudes characteristics are analysed. It is shown that wave propagation in a numerical model resembles propagation in a periodic medium. Consequently, Lamb wave propagation close to numerical Brillouin zone is investigated and characterized.
Analysis of elastic wave propagation in nonlinear media has gained recent research attention due to the recognition of their
amplitude-dependent behavior. This creates opportunities for increased accuracy of damage detection and localization,
development of new structural monitoring strategies, and design of new structures with desirable acoustic behavior (e.g.,
amplitude-dependent frequency bandgaps, wave beaming, and filtering). This differs from more traditional nonlinear analysis
approaches which target the prediction of higher harmonic growth. Of particular interest in this work is the analysis of
amplitude-dependent shifts in Lamb wave dispersion curves. Typically, dispersion curves are calculated for nominally linear
material parameters and geometrical features of a waveguide, even when the constitutive law is nonlinear. Instead, this
work employs a Lindstedt - Poincare perturbation approach to calculate amplitude-dependent dispersion curves, and shifts
thereof, for nonlinearly-elastic plates. As a result, a set of first order corrections to frequency (or equivalently wavenumber)
are calculated. These corrections yield significant amplitude dependence in the spectral characteristics of the calculated
waves, especially for high frequency waves, which differs fundamentally from linear analyses. Numerical simulations
confirm the analytical shifts predicted. Recognition of this amplitude-dependence in Lamb wave dispersion may suggest,
among other things, that the analysis of guided wave propagation phenomena within a fully nonlinear framework needs to
revisit mode-mode energy flux and higher harmonics generation conditions.
Acoustic Emission phenomenon is of great importance for analyzing and monitoring health status of critical structural components. In acoustic emission, elastic waves generated by sources propagate through the structure and are acquired by networks of sensors. Ability to accurately locate the event strongly depends on the type of medium (e.g. geometrical features) and material properties, that result in wave signals distortion. These effects manifest themselves particularly in plate structures due to intrinsic dispersive nature of Lamb waves. In this paper two techniques for acoustic emission source localization in elastic plates are compared: one based on a time-domain distance transform and the second one is a two-step hybrid technique. A time-distance domain transform approach, transforms the time-domain waveforms into the distance domain by using wavenumber-frequency mapping. The transform reconstructs the source signal removing distortions resulting from dispersion effects. The method requires input of approximate material properties and geometrical features of the structure that are relatively easy to estimate prior to measurement. Hence, the method is of high practical interest. Subsequently, a two-step hybrid technique, which does not require apriori knowledge of material parameters, is employed. The method requires a setup of two predefined clusters of three sensors in each. The Lamb wave source is localized from the intersection point of the predicted wave propagation directions for the two clusters. The second step of the two-step hybrid technique improves the prediction by minimizing an objective function. The two methods are compared for analytic, simulated and experimental signals.
The paper deals with the problem of Lamb waves dispersion curves sensitivity to the change of elastic constants
in composite materials. The framework of the present work is a more general problem of material constants
identification in thin plates made of composite materials. The approach is based on the analysis of guided waves
propagation and the related dispersion curves to find the underlying material elastic constants. In present work a
numerical study is performed to identify measurement directions and wave propagation modes that are most sensitive to
the change of the particular elastic constants. This approach will allow to optimize the material constants identification
procedure and experimental setup by specifying the preferred measurement directions and wave propagation modes. The
approach can be used within the Structural Health Monitoring framework to monitor material degradation of plate-like
structures made of composite materials.
The last few decades have seen a significant increase in research interest related to nonlinearities in micro-cracked and cracked solids. As a result, a number of different nonlinear acoustic methods have been developed for damage detection. The paper investigates nonlinear crack-wave interactions used for damage detection in plate-like structures. Semi-analytical modelling is used to investigate wave propagation in the vicinity of the crack. The focus is on non-classical crack model leading to wave modulations. Various physical phenomena (including fluctuation of temperature gradient) associated with these modulations are investigated. The work presented can be used for better understanding of nonlinear crack-wave interactions that are used for damage detection in structural health monitoring applications.
Two local computational strategies for modeling elastic wave propagation, namely the Local Interaction Simulation
Approach (LISA) and Cellular Automata for Elastodynamics (CAFE), are compared and contrasted in analyzing bulk
waves in two-dimensional nonlinear media. Each strategy formulates the problem from the perspective of a cell and its
local interactions with other cells, leading to robust treatments of anisotropy, heterogeneity, and nonlinearity. The local
approach also enables straight-forward parallelization on high performance computing clusters. While the two share a
common local perspective, they differ in two major respects. The first is that CAFE employs both rectangular and
triangular cells, while LISA considers only rectangular. The second is that LISA appeared much earlier than CAFE
(early 1990’s versus late 2000’s), and as such has been developed to a much greater degree with a multitude of material
models, cell-to-cell interactions, loading possibilities, and boundary treatments. A hybrid approach which combines the
two is of great interest since the non-uniform mesh capability of the CAFE triangular cell can be readily coupled to
LISA’s rectangular grids, taking advantage of the built-in LISA features on the uniform portion of the domain. For
linear material domains, the hybrid implementation appears straight-forward since both methods have been shown to
recover the same equations in the rectangular case. For nonlinear material domains, the formulations cannot be put into a
one-to-one correspondence, and hybrid implementation may be more problematic. This paper addresses these differences
by first presenting the underlying formulations, and then computing results for growth of a second harmonic in an
introduced bulk pressure wave. Rectangular cells are used in both LISA and CAFE. Results from both approaches are
compared to an approximate, analytical solution based on a two-scale field representation. Differences in the LISA and
CAFE computed solutions are discussed and recommendations are made for a follow-on hybrid implementation.
Although damage detection using Lamb waves has been investigated for many years, real engineering applications are
limited due to practical aspects related to implementation. Temperature effect is one of the major problems. It is well
known that temperature variations influence Lamb wave propagation response parameters. In practice it is important to
compensate for this effect. Experimental tests are often required to understand how temperature influences wave
propagation. Numerical simulation can ease this task preventing many time-consuming experiments. Simulated Lamb
wave responses can be used to develop new methods for temperature compensation.
The effect of temperature variations on piezoceramic transducer responses is investigated using finite element modelling.
The model takes into account temperature-dependent physical properties of low-profile PZT transducers and transducer
bonding layers. The model is used to predict the S0 and A0 Lamb response in aluminium plate for the temperature range
from -60 to +40°C. The study shows relevant changes in Lamb wave amplitude response caused by temperature
fluctuations. This approach can provide the basis for temperature compensation in ultrasonic guided wave damage
detection systems used for structural health monitoring applications.
Elastic waves are frequently employed in Non-Destructive Testing and Structural Health Monitoring systems for damage
detection and evaluation. Accurate and fast simulation of structural responses - due to high frequency ultrasonic
excitation - is possible by the application of graphical processing units (GPUs). The research work presented in this
paper is motivated by high demands for model sizes and GPU memories. The paper presents new developments related
to the simulation framework for elastic wave propagation. The proposed software environment uses the modified local
interaction simulation approach and graphical processing units to model ultrasonic wave propagation in structures. The
method has been enhanced by the application of multi-GPU workstations. This allows for effective simulations that can
be performed in complex structural components. The multi-GPU architecture and decomposition approach are
accompanied by absorbing boundary conditions for the model, enabling the reduction of model size and replacing
missing parts of the model by infinite space.
In this paper we present a new methodology for theoretical, numerical and experimental investigations of various
2D arrays' topologies. The theoretical evaluation is performed using frequency-dependent structure transfer
function that affects propagation of Lamb waves (LWs) through the dispersive medium and enables investigation
of the arrays' performance for a defined excitation signal. The numerical simulations are conducted using
local interaction simulation approach (LISA) implemented on the NVIDIA R CUDA R graphical processing unit
(GPU), which considerably accelerates 3D simulations of LWs propagation in a short time period. Finally,
scanning laser vibrometer is used to sense the LWs excited by PZT transducers, in multiple points corresponding
to the locations of the 2D array elements. In this way performance of various array architectures in the reception
mode can be evaluated experimentally without the need of physical prototype - a change of topology requires
only straightforward modification of the measurement points' distribution at the tested plate.
Guided ultrasonic waves are widely used in Structural Health Monitoring applications for inspections of large plate-like
structures. Wave propagation phenomena associated with guided ultrasonic waves are difficult to model for complex
engineering structures. Various simulation algorithms used in practice are not accurate and very expensive
computationally. The paper demonstrates new parallel computation technology offered by modern Graphics Processing
Units (GPUs) and Compute Unified Device Architecture (CUDA) used in low-cost graphical cards available in standard
PCs. Such systems enable calculations of very large models in minutes. The Local Interaction Simulation Approach
(LISA) algorithm have been implemented and used for wave propagation modelling. Application examples are related to
structural damage detection. The results demonstrate good accuracy and effective computational performance.
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.