Tuesday 18 September
Opening Plenary Paper – Roy Sharpe Prize Lecture
The emergence of thermographic NDE
Professor D Almond
The University of Bath
The lecture will outline the operating principles, capabilities and applications of thermographic NDE techniques. The techniques that will be covered will include: pulsed transient thermography, lock-in thermography and thermosonics (sonic IR or vibro-thermography).
Session 1A – Ultrasonic Transducers and Arrays
Improved transducer performance with new piezoelectric materials for non-destructive testing
M F Wallace1, S Cochran1, P Marin2, M P Walsh2, R Wright3 and R Marsh3
1Microscale Sensors, University of Paisley, Department of Engineering & Science, Paisley, PA1 2BE
2Piezo Composite Transducers Ltd, Crombie Lodge, Balgownie Drive, Aberdeen, AB22 8GU
3Tritech International Ltd, Peregrine Road, Westhill, Aberdeen, AB32 6JL
Piezoceramic-polymer composites have found increasing acceptance in applications in NDT. However, significant progression in piezoelectric materials has been made in recent years with the emergence of new higher performance single crystals. These new materials offer major improvements in transducer bandwidth and sensitivity, particularly when they are used in piezocrystal – polymer composite structures[1].
The work reported here presents the conclusions from a three-year investigation into the optimum transducer design process and practical limitations of PMN-PT single crystal. A comparison of piezocrystal-polymer composite transducers with conventional piezoceramic polymer composite transducers has shown the theoretical increases in performance with PMN-PT can be realised in practice. However, assessment of material uniformity[2] and changes in behaviour with temperature have indicated that present commercial adoption of single crystal is dependent on a reliable material supply and that tolerance to material variation is crucial to the design process. Furthermore, the price of these materials suggests that appropriate applications will require careful choice.
[1] P Marin-Franch, I Pettigrew, M F Parker, K J Kirk & S Cochran, 'Piezocrystal-polymer compostes: new materials for transducers for ultrasonic NDT', Insight Vol 46 (11), 2004.
[2] M F Parker, S Cochran, P Marin-France, D Choi & M P Walsh, 'Material property variation as a factor in commercial adoption of piezocrystals for composite transducer manufacture', Proc. IEEE 2004 Ultrasonics Symposium, (2005).
Below resonance operation of thin film ultrasonic transducer for NDT application in 800 kHz to 5 MHz frequency range
Jocelyn Elgoyhen, Katherine J Kirk, John Paul Hood, David Hutson
Microscale Sensors, School of Engineering and Science, University of Paisley, High Street, Paisley,
PA1 2BE, Scotland UK. T: 0141 8483428; E: jocelyn.elgoyhen@paisley.ac.uk
Thin film piezoelectric materials like aluminium nitride (AlN) are increasingly used for high frequency ultrasonic applications in film bulk acoustic resonator (FBAR) devices. In those implementations the thin film transducers are expected to operate at around their thickness mode resonance frequencies which is commonly in a range between 700 MHz to over 1.5 GHz.
Our work[1] has shown that a 4 µm thick piezoelectric AlN film on a metallic substrate can be used as an ultrasonic transducer well below its resonance frequency. Recent experimental measurements have showed that ultrasonic pulses with fundamental frequency between 800 kHz and 5 MHz could be used for pulse-echo and through-transmission measurements on aluminium, stainless steel and ferritic steel substrates. This may open the way for NDT applications using thin film transducers.
This paper investigates the mechanism behind the generation of ultrasonic wave below the resonance frequency of the transducer. SPICE simulation and equivalent circuit analysis are used to shows how in such circumstances the transducer thickness is not the predominant factor that controls the frequency of the acoustic wave. Instead it is mostly dictated by the harmonic content of the excitation pulse that reaches the transducer. The aim of this work is to engineer the electrical side of such system to control the frequency of operation of a thin film transducer.
[1] J P Hood et al, BINDT NDT 2006 Conference, 11-14 Sept 2006, Stratford-upon-Avon, UK.
Couplant-free thin film transducers for high temperature ultrasonic NDT
John Paul Hood, Katherine J Kirk, Jocelyn Elgoyhen, David Hutson
Microscale Sensors, School of Engineering and Science, University of Paisley, High Street, Paisley,
PA1 2BE, Scotland UK. T: 0141 848 3530; E: hood-ph0@wpmail.paisley.ac.uk
Aluminium nitride is a compound with piezoelectric properties and a high Curie temperature, which implies it could have applications in ultrasonic high temperature non-destructive testing (NDT)[1,2]. In addition it can be deposited directly onto components and therefore does not require couplant.
In order to allow NDT measurements on high temperature plant to be carried out while the plant is online there are certain specifications the thin film transducer must meet. The transducer must be able to work over a suitable frequency range, be able to be deposited onto industrial materials, and survive plant operating temperatures.
A study has been carried out which shows that these specifications can be met. A thin film transducer (4 µm thick) deposited on to a steel coupon (1 cm thick) has been successfully used with a Krautkramer USK7 flaw detector. A device made on a ferritic steel steam pipe section (65 mm thick) with a ground surface has been used in both pulse-echo and through-transmission modes at 5MHz. High temperature tests have also shown the transducers to work up to 600°C.
[1] J P Hood et al, BINDT NDT 2006 Conference, 11-14 Sept 2006, Stratford-upon-Avon, UK.
[2] I Atkinson et al, Eighth International Conference on Creep and Fatigue at Elevated Temperatures, July 22-26, 2007, San Antonio, Texas.
Applications of new piezoelectric materials for high performance ultrasonic NDT transducers
M F Wallace1, P Marin2, H Mulvana3, M P Walsh2, T W Button4, H Hughes4, J Elgoyhen1, D MacLennan3 and S Cochran1
1University of Paisley, Paisley, Scotland, UK, 2PCT Ltd, Bridge of Don, Aberdeen, Scotland, UK
3University of Strathclyde, Glasgow, Scotland, UK, 4Applied Functional Materials Ltd, Birmingham, England, UK
Piezoelectric material is the most common means to transform electrical excitation signals into ultrasound and ultrasound echoes back into electrical received signals in ultrasonic transducers and arrays. In most such devices for non-destructive testing (NDT) the piezoelectric material is either lead zirconate titanate (PZT) or a PZT-polymer composite. However, there are significant opportunities to obtain higher performance for specific applications using different or more novel materials and several instances of this are discussed in the present paper.
- Very high spatial resolution testing is important for materials such as titanium metal matrix composites, for surface coatings and for miniature components. Existing high frequency transducers are often based again on LNO or on polyvinylidene fluoride (PVDF). However, a much higher performance alternative may become available through the development of microscale PZT-polymer composite material.
- Testing at elevated temperature is an increasingly widely recognised problem. Here, possible solutions are based on lithium niobate (LNO) or aluminium nitride. More exotic solutions such as LNO-cement composites have also been investigated.
- There is significant interest in the possibility of surface conformability to allow ultrasonic arrays to couple efficiently to non-planar surfaces. The only commonly recognised piezoelectric material that is inherently flexible is PVDF but it has very poor sensitivity. Another alternative that is emerging increasingly is flexible PZT-polymer composite.
- The performance of existing NDT transducers may be inadequate for testing difficult materials such as cast iron and austenitic stainless steels which have relatively granular structures. In this case, the additional sensitivity of new piezocrystals such as lead magnesium niobate doped with lead titanate (PMN-PT) may be of significant value.
Ultrasonic Arrays: a comparison between medical and NDE requirements
Paul D Wilcox and Bruce W Drinkwater
RCNDE, Department of Mechanical Engineering, University of Bristol, Bristol, BS8 1TR, UK
Ultrasonic arrays have been used for NDE applications since the 1980s but their use for medical imaging goes back to the 1950s. Historically, the medical market has been the main driver in the development of array transducers and array instrumentation technology, with the NDE sector adapting the technology to its own applications. This paper addresses the question of designing and optimising arrays and image processing specifically for NDE applications.
There are two key differences between medical and NDE array applications. Firstly, in medical applications the target is non-stationary and it is therefore essential to obtain a high image frame rate both to avoid motion blur and to observe phenomena such as the human heartbeat. Hence while a high frame rate may be desirable in NDE it is a necessity in medicine. A second difference is that NDE images are dominated by signals from a relatively small number of strong scatterers and there is a high acoustic contrast between defects (e.g. cracks) and the parent material. In comparison, medical images are characterised by low amplitude scattering, observable throughout an image. Taken together these differences mean that NDE requirements are not best met by simple transfer of medical technology. For example, the concept of capturing and post processing the full matrix of raw data (ie the time-trace from every possible transmitter receiver combination) is appealing in NDE and too slow for the medical field. Using this approach, the best possible imaging resolution can be obtained by focusing an array on every imaging point in both transmission and reception. Improved reflector characterisation can be achieved by considering the array as a number of sub-apertures to determine reflector orientation and shape. The full matrix of data also provides the information necessary to apply super-resolution imaging algorithms that allow the wavelength diffraction limit for resolution to be overcome.
Session 2A – Phased Arrays
An insight into the behaviour of periodic piezoelectric composite array transducers incorporating contemporary constituent materials
G Harvey, R L O'Leary, A Gachagan and G Hayward
The University of Strathclyde, Glasgow, Scotland, UK
Ultrasonic transducer arrays are well established in both biomedicine and sonar systems and are attracting significant interest within the field of NDE. Across this diverse application range, the piezoelectric ceramic composite configuration is commonly used as the active material, where selection of the constituent materials is crucial in the design process. New advances in both piezoelectric and polymeric materials have been shown to improve array sensitivity, bandwidth and imaging performance in biomedical and sonar applications. This work investigates the potential improvements in NDE array performance which may be realised through the application of these new materials.
This paper describes a theoretical investigation into the influence of the constituent materials on periodic piezoelectric ceramic composite array transducer performance. Finite element (FE) modelling has been employed to predict the behaviour of wedge coupled array transducer configurations operating into a steel component. The performance enhancement offered by new single crystal peizoelectric materials is compared to standard PZT-based configurations. Moreover, a novel passive polymer material, possessing low longitudinal loss and high shear loss, is shown to significantly reduce inter-element mechanical cross talk across the array aperture. Finally, the imaging performance of piezocomposite array transducers incorporating these new materials is evaluated using A-scan predictions from simulated defects.
Development of an efficient conformable array structure
G Harvey, J Mackersie and A Gachagan
Centre for Ultrasonic Engineering (CUE), Dept of Electronic and Electrical Engineering, University of Strathclyde, Glasgow, Scotland, UK.
NDT inspection of complex component geometries would benefit greatly from an efficient flexible piezoelectric material. Additionally, utilising a 2D array configuration can provide enhanced coverage map and defect detection capability of an inspection process compared with linear phased arrays. This paper describes a piezoelectric composite array structure that combines the performance advantages of a conventional 1-3 piezocomposite with the increased flexibility associated with the piezo-platelet arrangement. Specifically, each array element in the 2D matrix is a fine scale piezocomposite structure, with a relatively large, flexible, polymer phase between each element. Both periodic (using diced piezoceramic) and random (incorporating piezoceramic fibre technology) piezocomposite array structures are investigated for application as the array elements. Experimental measurements of electrical impedance, surface displacement and mechanical cross talk in these structures demonstrate good correlation with finite element results obtained by employing the PZFlex code. Moreover, the use of a random piezocomposite structure under each array element does not degrade the transducer sensitivity and bandwidth performance and actually provides superior conformability over corresponding periodic designs.
A modular FPGA-based ultrasonic array system for applications including non-destructive testing
S Triger1, J Wallace1, L Wang1*, S Cochran2, J Saillant2, F Afroukh2 and D R S Cumming1
1University of Glasgow, Glasgow, Scotland, UK
2University of Paisley, Paisley, Scotland, UK
1*Formerly at University of Glasgow, now at Imperial College, London, UK
This paper reports work aimed at the development of an ultrasonic imaging system which can be customised for multiple applications, including and within non-destructive testing (NDT), by tessellating in-system programmable modules. Each module can be thought of as a tile containing an individual ultrasonic array, and multiple tiles can be considered capable of assembly into an mosaic. The ability to form a larger array as a mosaic of physically identical building blocks has the potential to make manufacturing easier and to cut costs whilst preserving flexibility through the capability to program the tiles to work together in ways suitable for specific applications. The key component is an autonomous module containing the ultrasonic array and all the electronics necessary to operate it. This contrasts with most previous research on system integration which has focused only on the transducer and front end electronics.
In the present work, a 4 x 4 element 2D piezoelectric array with a 16 mm x 16 mm aperture has been produced, with the entire transmission and reception electronics within the same footprint. This permits the desired tessellation of multiple tiles to form larger arrays. The proximity of the transducer array and electronics removes the need for cabling, reducing signal degradation due to cross talk and interference. In addition, it avoids the problem of electrical impedance matching of cable between the array elements and the electronics. Any of a tile’s 16 array elements can be individually excited with 10 ns timing resolution by user-defined coded excitation waveforms generated by a field programmable gate array, maximising application flexibility. Presently, the elements are excited with modified Barker codes to achieve acceptable signal to noise ratio (SNR) even with excitation signal amplitudes as small as +/- 3.3 V. Signals received from all 16 elements are input to individual preamplifier stages whose outputs are sampled following time gain compensation and digitised with 12 bit resolution for initial signal processing, including decoding, before transmission to a host PC which allows the signals to be visualised.
Pulse-echo insertion loss of 40 dB has been measured from back wall reflections in 75 mm thick aluminum without decoding, and results with decoded signals show adequate SNR with +/-3.3V excitation at an operating frequency of 1.21MHz, within the range required for deep penetration in nuclear power plant. Crucially, the ability to construct 2D arrays of any size and shape from generic building blocks represents a departure from almost all previous work in ultrasound, which has traditionally been highly application specific. This may allow ultrasonic NDT to be used in applications for which the investment in customised devices could not previously be justified.
Correlation of ultrasonic phased array inspection and fatigue performance of FSW joints
D S Caravaca
TWI Ltd
This paper is based upon a TWI Core Research Project for the development of Friction Stir Welding (FSW) in aluminium.
FSW of aluminium alloys is now an established joining technique and there is increasing application of it to joining of critical components and structures. There is a requirement to develop inspection techniques for flaws that may be generated, when welding conditions deviate from the optimum. These flaws can be volumetric (voids), joint line flaws (JLF), such as lack of bonds, or joint line remnants (JLR) that are unique to this welding technique. The JLR is a region of weakly bonded material in the root of the weld and it is extremely difficult to detect with any non-destructive testing method due to its transparency to mechanical or electromagnetic waves. A procedure using a phased array technique has been developed to detect these flaws. This procedure, developed from that described by Kleiner and Bird(1), is based on the calculation and comparison of the average value of the grain noise in both weld and parent plate. Hence, instead of detecting the defect directly, a characteristic pattern of the noise in the weld root region is measured, thereby determining whether the weld nugget has been correctly forged. To demonstrate the effectiveness of this method, the indicated flaw sizes and fatigue life were compared, and they show a strong relationship.
Session 2A(2) – Advanced Ultrasonics
Application of Total Focusing Method in immersion ultrasound non-destructive evaluation
Gerald Harvey1, Andrew Tweedie2, Caroline Holmes3, Paul D Wilcox3, Richard L O'Leary1 and Anthony Gachagan1
1Centre for Ultrasonic Engineering, Department of Electronic and Electrical Engineering, University of Strathclyde, Glasgow, G1 1XW, UK.
2Alba Ultrasound Ltd, Unit 1, Building 3, Todd Campus, West of Scotland Science Park, Glasgow, G20 0XA, UK.
3Department of Mechanical Engineering, University of Bristol, Bristol, BS8 1TR, UK.
2-D ultrasonic arrays can allow for rapid scanning of components and when coupled with full matrix data capture and the total focusing method (TFM) can provide improved image resolution and inspection coverage.
This paper describes the use of a 50x44 mm aperture, fully populated 2-D piezoelectric composite array with a 2mm element pitch operating at 2 MHz in immersion mode. An aluminium test block containing a series of 1 mm slot defects with varying degrees of skew to the normal was interrogated. The full matrix of data from this 300 element array was captured and a 3-D TFM algorithm implemented in MATLAB for the offline post-processing of the data. Moreover, the algorithm was modified to accommodate the refraction effects at material boundaries in this immersion scanning configuration, and to support vector-based TFM for determination of defect orientation.
The resultant images show that TFM can be applied to full matrix 2-D array data to provide a rapid, volumetric immersion scanning system, with improved resolution over conventional B-mode imaging. Furthermore, the improved SNR achieved through TFM may make immersion scanning an attractive alternative to contact probes in certain situations.
Bats and coded ultrasonic sequences – the potential implications for improved NDE of structures and materials
Gordon Hayward, John Soraghan, Gareth Pierce, Frederic Devaud and Najaf Akbar
The Centre for Ultrasonic Engineering, The University of Strathclyde, Glasgow G1 1XW.
Research into the behaviour of mammalian creatures such as bats and cetaceans has increased significantly in recent years, with the primary driver being the constant search for resolution enhancement, mainly in ultrasonic imaging systems. The implications for NDE are potentially very exciting, with applications ranging from navigation and guidance of robotic vehicles, through to sub wavelength resolution in ultrasonic array imaging applications.
This paper will describe and present results from a major programme of research that is currently underway at Strathclyde. A signal processing algorithm, derived from analysis of the neural system of the FM bat, is described and evaluated on a range of ultrasonic data. The algorithm is shown to discriminate between targets separated by distances much less than the emitted ultrasonic wavelengths and its ability to operate successfully on data that has low signal to noise ratio and that can be limited in spectral content, is demonstrated.
The types of coded signals used by FM bats are discussed and potential applications for the new processing methodology in the world of NDE are described. These include imaging using ultrasonic arrays and guidance systems for micro-miniature robotic systems, in which groups of mobile sensing agents can inspect a structure autonomously.
Characterising the anisotropy of a welded high-Nickel alloy to assess the feasibility of ultrasonic inspection
G Elston
Doosan Babcock Energy Ltd
Alloy 263 is an austenitic high-Nickel alloy being considered for use in the super-critical steam cycle of the next generation of power stations. As part of the qualification and approval process, it is essential to demonstrate that components made of Alloy 263 can be inspected with standard NDT techniques, especially ultrasonic testing. We report on our investigations of the ability of ultrasonic techniques to inspect welded Alloy 263 components.
Welded austenitic materials are anisotropic in their ultrasonic properties. The speed of sound and the attenuation vary as a function of the angle of incidence with respect to the principle grain axis, which means they are transversely isotropic. Sound energy propagating at an angle to the principle grain axis is deviated towards the higher sound speed directions by a process analogous to refraction at a boundary between two different materials. Measurements of the deviation and attenuation of ultrasonic beams directed through cuboid blocks of weld material provide data for comparing the ultrasonic properties of different materials.
Stainless steel is a commonly welded austenitic material that is routinely inspected using ultrasonic techniques. Equations exist that adequately describe its anisotropy and we have empirical knowledge of the normal range of its attenuation values. It is this combination of theoretical and empirical values that are compared to the experimental data from Alloy 263.
We have begun a series of experiments to fully characterise the ultrasonic properties of Alloy 263 in both parent and welded states. Early indications are that welded Alloy 263 exhibits less anisotropy and lower attenuation than stainless steel; we therefore have a good indication that Alloy 263 will be amenable to ultrasonic inspection. Once this characterisation is complete we will be able to machine representative defects into a sample of welded Alloy 263 pipe and perform a realistic ultrasonic inspection.
