 |  |
In this section:
|
Structural Damage Detection Using Lamb Waves and Piezoceramic Sensors |
Lamb waves are dispersive plate waves that occur for traction-free forces on both surfaces of the plate. The velocity of these waves depends on the product of frequency of excitation and thickness of the plate. They can propagate long distances and are used for damage detection of plate-like structures. Lamb waves are the most widely used guided waves for damage detection. The strategy of monitoring is extremely important for successful damage detection. The basic factors, which determine the Lamb wave based damage detection analysis are related to properties of the structure under inspection, transducer schemes, choice of excitation input signal, and appropriate signal processing. Various methodologies have been developed for successful damage detection in metallic and composite structures. This includes damage features which do not depend on loading and environmental conditions.
|
 |
Aluminium panel with piezoceramic sensors. |
|
 |
|
Lamb wave response from piezoceramic sensors |
Structural Damage Detection Using Lamb Waves and Laser Vibrometry |
Actuation and sensing of Lamb waves can be accomplished using various types of transducers. A significant number of transducers is required in practice for monitoring of large structures. This is often not possible or acceptable. A portable scanner, i.e. a pair of transducers that can crawl on the surface is one possible solution. The difficulty is the need for liquid couplant. More recently dry coupling methods, using non-contact electromagnetic acoustic transducers and air-coupled transducers have been applied in NDT for scanning large surfaces. However these transducers have very low sensitivity due to the large acoustic impedance mismatch between air and solid materials and due to the high attenuation of ultrasound in air for high frequencies. An alternative approach is to use lasers. Laser vibrometer has been used for damage detection in metallic structures. The method allows for non-contact measurements of local amplitude variations that cannot be achieved with traditional transducers. A full Lamb wave field has been obtained allowing for structural damage detection.
|
 |
 |
|
 |
|
Damage detection results based on laser vibrometry |
Structural Damage Detection Using Acousto-Ultrasonics |
Various acousto-ultrasonic techniques have been developed for structural health monitoring. These techniques are based on stress waves introduced to a structure by a probe at one point and sensed by another probe at a different position. The frequency of these waves can go up to MHz range. This high-frequency excitation results in a large number of mixed modes. Transducer responses include not only directly propagating wave modes but also reflected and scattered modes. This rich frequency content requires advanced signal processing techniques for damage detection. The method can be applied for crack monitoring in metallic structures and impact damage detection in composite structures. Various types of transducers can be employed for monitoring. This includes piezoceramic sensors, which can become an integral part of a monitored structure.
|
 |
Crack monitoring using acousto-ultrasonics |
|
 |
|
Damage detection results |
3-D Laser Vibrometry for Structural Damage detection |
Lamb wave inspection uses guided ultrasonic waves to detect damage in structures. Using a new detection technology known as 3-D Scanning Laser Vibrometry, structural damage is clearly identified by locally increased in-plane and out-of-plane vibrations. 3-D laser vibrometer scans can reveal structural damage and its severity such as crack length and delamination area. Simple contour maps and profiles of Lamb wave amplitude across the structure are sufficient to see the damaged areas. The method does not involve studies of complex Lamb wave propagation in the structures, baseline reference measurements in undamaged structures, or signal post-processing to extract damage-related features. The method is straightforward, fast, reliable and immune to environmental effects. |
 |
 |
|
|
|
Damage Detection in Concrete Structures Using Smart Sensor Technologies |
Non-destructive techniques using guided ultrasonic waves have been applied to monitor dynamics elastic constants of concrete and detect internal defects in concrete structures. PZT patches have been bonded on the surface of concrete blocks and used to generate and receive guided ultrasonic waves. Various signal processing parameters can be correlated with internal cracks in concrete during loading.
|
 |
 |
|
Damage detection in concrete structures |
Impact Damage Detection Using Piezoceramic Sensors |
Work in this area has focussed on passive and active damage detection based on piezoceramic sensors. In the passive mode, sensors are bonded or embedded in structures in order to monitor impact strain data. The strain data are then used for detection and location of impacts. The study involves advanced signal processing procedures based on pattern recognition. The active approach is based on acousto-ultrasonic waves introduced to structures. Applications in this are include Lamb wave inspection techniques.
|
 |
Experimental facilities for impact tests in composite materials |
Acousto-Ultrasonic Sensing Using Fibre Bragg Gratings |
The aim of this project is to to develop and implement optical fibre Bragg grating sensors for acousto-ultrasonic based structural helath. This state-of-art sensors will be capable to measure strains, temperature and monitor damage at the same time. Multi-functional sensors have a great potential for structural health monitoring incolving smart technologies. Advanced signal processing, including data fusion, is an important part of this project. The work is part of the joint-location PhD project in collaboration with the DaimlerChrysler Research Centre in Ulm, Germany and Strathclyde University in Glasgow, UK.
|
 |
Responses from the optical fibre Bragg-grating sensors. |
Nonlinear Accoustics |
Nonlinear effects in acoustic signals are widely used for structural damage detection. Application examples include methods based on acoustic/structural wave propagation and utilise various nonlinear phenomena such as generation of higher frequency harmonics, frequency mixing or frequency modulations. The work in the group is focused on the application of low-profile piezoceramic transducers for vibration excitation used in nonlinear acoustics. The method, used for fatigue crack detection in metallic structures, shows more sensitivity to damage than any other technique based on linear ultrasonic effects.
|
 |
 |
New Damage Detection Technique |
Recent work in the Department has resulted in a new method of structural health monitoring of structures. The technique provides a unique and extremely effective method of crack detection and monitoring which is more robust and accurate than current methods available. It enables damage parameters to be directly related to fatigue analysis and crack length. European and US patents have been applied for and it is clear that this invention has significant commercial possibilities.
|
 |
 |
Acousto-Ultrasonic Wave Propagation Modelling |
Knowledge and understanding of of wave propagation can ease the interpretation of damage detection results based on acousto-ultrasonic inspection. Various numerical tools based on the local interaction simulation approach (LISA) have been developed for modelling of Lamb wave propagation and wave interaction with defects in metallic structures. The LISA modelling technique is well suited for wave propagation in complex materials with different physical properties and discontinuities This is particularly important when piezoceramic sensors and/or structural defects are involved in modelling. The study involves various monitoring strategies based on different sensor locations and environmental conditions.
|
 |
2-D Lamb wave propagation in an aluminium plate |
|
 |
|
LISA wave propagation modelling software |
Optimal Sensor Location Procedures |
Work in this area has focussed on various methods based on combinatorial optimisation, neural networks and mutual information. the active control of structural vibrations. We have developed modelling procedures that involve finite element model updating to create model-based controllers capable of suppressing structural vibrations without introducing spillover effects. Constrained layer damping is typically used in conjunction with active piezo-electric elements.
|
 |
Optimal sensor location problem in a composite SMART panel |
Dynamics of Structures and Machines: |
Modal Analysis |
The analysis of dynamic properties of structures and machines structural vibration under vibration excitation is important in many engineering applications. Various types of vibration excitation are used in practice. This includes: broadband, sweep sine and impulse excitation. The study can reveal structural modes, resonances and damping levels. Often these experimental results are correlated with Finite Element Analysis. The work carried in Sheffield is important for Nonlinear Acoustic applications.
|
| ||||||
|
| |||||||
Dynamics of Structures and Machines: |
Vibration Analysis using the Hilbert-Huang transform |
Many engineering systems exhibit time-variant behaviour due to varying physical parameters and/or nonlinearities. The analysis of such systems can be eased with the application of the Hilbert-Huang transform. The method consists of two major procedures: the Empirical Mode Decomposition (EMD) and the Hilbert Spectral Analysis (HSA). The former is used to decompose any multi-component signal into mono-component signals, often called intrinsic mode functions. The later is can be applied to obtained signal instantaneous frequencies. Although the EMD procedure is empirical the intrinsic mode functions are physically meaningful. The instantaneous frequencies associated with these functions can be used to estimate varying physical and/or modal parameters time-variant an/or non-linear systems.
|
|
||||||
|
| |||||||
Advanced Signal Processing: |
Pattern Recognition |
There exist three different approaches to pattern recognition. These are statistical, syntactic and
neural based methods. A number of neural network methods have been developed for pattern recognition. Most applications are related to signal feature mapping for damage detection.
|
 |
|
|
Advanced Signal Processing: |
Genetic Algorithms |
Genetic Algorithms are optimisation search procedures simulating natural evolution. Possible candidate solutions are coded as chromosomes. Random operations based on natural selection are used to evolve the initial population. This approach combined with neural networks has been used successfully for optimal sensor location problems in damage detection and for feature extraction/selection applications in data compression algorithms.
|
 |
|
|
Simulated Annealing |
The algorithm of simulated annealing aim is an optimisation search method based on an analogy to thermodynamics. The process of annealing and the Boltzman theorem of thermal equilibrium are used to find an optimal solution in a complex search space. The method has been used successfully for ridge extraction from the wavelet transform. This procedure is part of the damping estimation technique in MDOF systems.
|
 |
|
|
Advanced Signal Processing: |
Damping Estimation Techniques |
A number of different methods of damping estimation for MDOF systems have been developed. The methods are based on the time-scale decomposition of the analysed system impulse response. The continuous wavelet transform is used to decompose the impulse response into the time-scale domain. The wavelet transform cross-sections, the impulse response recovery procedure based on wavelet filtering and the wavelet ridge detection procedure are used for damping estimation. The methods have been applied to simulated and real experimental data.
|
 |
|
|
Wavelet-Based FRFs |
The system response to load excitation can be represented using classical FRFs. In general these characteristics are not adequate to establish the proper input/output relationship of the system with strong non-linearities and varying physical properties. The classical FRF has been extended using wavelet and cross-wavelet densities instead of spectral densities. The method has been applied successfully to simulated and experimental vehicle vibration data.
|
 |
Wavelet-based FRF for a non-linear system |
Analysis of Non-linear Systems |
All physical systems exhibit in practice non-linear behaviour. New-procedures of non-linear system identification have been developed in the Department. These procedures employ slowly-varying, time-dependent amplitude and phase functions of the impulse response of the system. These instantaneous characteristics are obtained from the ridges and skeletons of the wavelet transform. The procedure can be used for MDOF systems.
|
 |
Cross-wavelet transform of the Duffing oscillator exhibits non-linear stiffness behaviour. |
Data Processing with Fractals |
Stationary periodic series are invariant under translations in the time domain. There exist many physical processes, which display invariance under scale change. Such signals have fractal waveforms. The fractal structure of these signals can be reveal using the wavelet transform. The method has been applied successfully to analyse coherent structures in boundary-layer data, chaos and noise.
|
 |
Devil’s staircase fractal. |
Optical Measurements and Wavelets |
High-speed measurements have been used in many areas of engineering applications. This includes vibration analysis where images can provide significantly more measurement responses than classical transducer-based analysis. Wavelet-based algorithms for edge detection extraction have been developed and applied in experimental modal analysis. The study involves the estimation of modeshapes, natural frequencies and damping parameters. The method has a potential not only for automotive applications but also for small-scale structures (e.g. MEMS) where transducer-based measurements are not possible.
|
 |
 |
Aircraft image and its wavelet-based edge. |
Data Compression for Vibration Analysis |
Various algorithms based on the orthogonal wavelet transform have been developed for data compression. Data compression in vibration analysis can be used not only for effective storage and transmission but also for feature extraction/selection, solving of partial differential equations and model reduction/synthesis techniques. The first application is particularly relevant to damage-oriented features in structural health monitoring. The second application involves compression of matrices associated with matrix-based complex equations. The last area of research is related to model reduction and synthesis of MDOF complex structures.
|
 |
 |
Wavelet–based data compression for modal reduction/synthesis of structures |
| |