IOP Groups
Modern Applications of Acoustic Emission
Cardiff University, January 16 2003
This one-day meeting was very successful, with 40 delegates and speakers attending, the vast majority of whom took part in lively discussions following each speaker and in the discussion on “Future Directions for AE” at the end of the day. The delegates were almost exactly divided between academia and industry and it was pleasing to note the extent of usage and knowledge of the AE technique in the UK.
Phil Cole, of Physical Acoustics Limited, gave an outline of the process industry applications of AE. These applications are mostly well-established in the industry and result in massive cost savings. An example of this is tank floor inspection for bulk storage tanks in the oil industry – it costs over £1M to remove a 100m tank from service and there are environmental issues of opening these tanks, disposal of waste is costly and there is a risk of local contamination. The TANKPAC technique uses sensors located around the outside of the tank and can detect both corrosion and leak. Examples of successful applications around the world were given. The role of AE in risk-based inspection was summarised with examples spanning many industries. The subsequent discussion included:
- the use of other types of corrosion sensor, which aren’t suitable as they are invasive and suffer from wear
- the method for achieving high temperature design, no soldering, different crystal etc, and the use of waveguides
- issues for 3D locations on complex geometries
Pete Theobald of the National Physical Laboratory introduced the work of NPL and the Acoustics programme in general, which is driven by industry. The AE sensor calibration method developed at NPL compares the AE sensor response with that of an interferometer measurement of the out-of-plane displacement of the surface. The source transducer is excited with tone bursts at discrete frequencies and the transmission medium is a glass block. The main measurement issue is the fact that this calibrations is performed using the direct compression (longitudinal) wave and not the Rayleigh wave, due to phase variation across the sensor face. The lowest useable frequency is 200kHz. The method allows an absolute comparison of sensor performance. NPL have also developed a point-contact transducer for use as an in-situ system performance verification, to replace the popular H-N (pencil lead) source and provide a traceable link for all AE measurement. FE modelling is being employed to investigate the OOP and IP displacements and issues include coupling, robustness and mechanical impedance. NPL are also developing a modelling facility, in conjunction with Fraunhofer Institute, to model the reference facility from excitation of the reference to output of the detecting sensor. The subsequent discussion focussed on the accepted drawbacks of the sensor calibration method, issues of traceability, transducer sizes and build up of crystal resonance due to use of the tone burst.
Keith Worden, of Sheffield University, treaded us to a well explained tutorial in classification techniques. He outlined the methods available, and problems associated with automatic classification of AE data. This included the use of Principal Component Analysis (PCA) for data dimension reduction. Keith demonstrated the successful identification of three main AE types within a set of test results containing AE from several sources, using the first two principal components. The main automatic classification techniques were explained, including various Gaussian statistical methods and Neural Network methods. It was demonstrated that the AE data reduced by PCA could be automatically classified into subsets representing different generation mechanisms using both statistical methods and Neural networks, but noted that more general AE problems may require more advanced data reduction or signal processing.
We were very grateful to Mike Blanche, from the Rutherford Appleton Laboratory, for stepping in at the last minute to provide an excellent account of their work under the recent AEGIS project, in conjunction with CRES, Delft University of Technology, Envirocoustics ABEE, Geobiologiki, Aerpac B.V., University of Patras, Cranfield University and Euro Physical Acoustics. This EC funded project combined the expertise of blade manufacturers, testing institutions, and acoustic emission consultants to develop a system for listening to and analysing the sounds of material failure at a microstructural level in order to give advanced warning of critical damage.
A static “load and hold” Acoustic Emission Examination Load (AEL) test was used for a wind turbine blade mounted on a wind turbine, using a pulley mechanism to load the blade. The retro-fitted loading device was able only to fully test the mid-span of the blade and not the more critical blade root area. Any widespread application of the technique would require the loading system to be built in to the tower and suitable load application points sited on the blade. The principles of a real-time AE system were demonstrated on the Windharvester wind turbine at RAL. A broadband radio transmission system was developed to transmit the AE data from the rotating frame to the ground without losing any signal resolution. A 112m metre diameter turbine, with a blade length of 49m, was instrumented with fibre-optic AE sensors and it was found that is possible to derive information relating to the location, type of damage, and even criticality.
Dr John (Iain) Steel, of Heriot-Watt University presented his work on the application of AE to reciprocating machines. This application provides for potential incorporation into engine-management systems. The research and field testing undertaken has covered an enormous variety of applications, including gas engines, diesel engines (power stations and ships) and compressors and has encompassed a variety of faults. In these applications, AE signals arising due to impacts, fluid/gas flow, sliding(wear), crack propagation and combustion. The faults considered include tappet clearance, valve faults, gasket leakage, oil condition, injector faults, fuel condition, liner scuffing and a variety of general engine operating parameters. AE signals arising from these faults and conditions were presented and explained. Some work on source location and attenuation in the engine block was presented and the potential use of Lamb wave analysis outlined. The main question following the presentation was whether AE can be used for any engine with any fault and the simple answer is no! However the potential of the technique has been demonstrated and clearly there is scope for much research in this field.
Tim Bradshaw, formerly of Cardiff University, now employed by Physical Acoustics Ltd, outlined some specialist applications of AE for damage detection. A large range of topics was covered: F-111 Cold Proof Test; VC-10 Proof Pressure Test; SAAB JAS-39 Gripen Full-Scale Static Test; AE Monitoring of the ARIANE IV SPELDA; In-Flight AE Monitoring of the DC-XA Rocket; Steel Box Girder Inspection; Concrete structure Inspection and Continuous Integrity Monitoring of Structures (AE Monitoring of an Offshore Structure). The aerospace applications are generally well accepted, providing an invaluable protection against damage during proof testing and locating onset of early damage. The DC-XA VTOL test demonstrator proved the ability for AE for in-flight structural integrity monitoring and low cost instrumentation. The civil engineering applications covered an established procedure, developed by PAL in conjunction with Cardiff University, for the inspection of steel box girders. This is now an accepted commercial procedure and has been applied successfully across much of the UK motorway network. Another successful civil engineering application is that of concrete structure inspection. Finally, the use of AE for continuously monitoring major structures was demonstrated by the monitoring of the Talisman oil platform, in the North sea. The presentation outlined the techniques and benefits of AE monitoring for damage assessment across a wide range of industrial applications and highlighted the reliability and cost effectiveness of the new generation of AE systems.
Finally, Tony Martin, of BAE Systems, explained all of the issues involved with getting AE systems onto aircraft. He explained why structural health monitoring is essential, the sources of AE in aircraft and the requirements for “the perfect sensor” and the integration of systems into the current avionic systems.. A key issue is that no one system can detect all types of damage. Another crucial point is the qualification of equipment and safety critical status. He discussed the potential areas for AE monitoring and the types of damage that can potentially be detected. The issues of sensor selection were discussed in detail, with the tentative conclusion that optical sensors are most likely to succeed if issue of sensitivity can be resolved. The issues of surface mounting or embedding sensors was discussed and the applications of other health monitoring systems. The conclusion is that provided that the benefits of fitting a damage detection system out weigh the costs associated with integrating, use of and qualification then such a system could be use on an aircraft. The final hurdle is ti integrate an SHM system into a flying (fast jet) test bed, which may occur in the near future. Questions included the problem of fatigue of “smart patch” sensors, the number of sensors required for full coverage, and the application of TOA algorithms for location.
The final session of the day was an open discussion, aimed at looking at the future for AE. It was hoped that industrial participants would outline their requirements and that practitioners would outline where technology was leading. The discussion actually highlighted the need for standards and better communication between researchers and it was agreed that an AE network would be beneficial.
^ To the top ^