A study of EMAT (electromagnetic acoustic transducer) operation on ferromagnetic metals

Ultrasound is widely used as a tool for non destructive testing, being completely non invasive and much safer than some other methods, X ray techniques for example. Its low relative cost is also attractive purely from a business standpoint. Its use, however, is restricted more to applications where...

Full description

Bibliographic Details
Main Author: Kiteley, Ian David
Published: University of Warwick 1999
Subjects:
Online Access:http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.343125
Description
Summary:Ultrasound is widely used as a tool for non destructive testing, being completely non invasive and much safer than some other methods, X ray techniques for example. Its low relative cost is also attractive purely from a business standpoint. Its use, however, is restricted more to applications where the ultrasonic transducer and test piece can be brought into close contact, thus allowing the use of couplant gels. In situations where a fully non contacting method is required the cost and complexity of an ultrasonic system is considerably higher. These usually take the form of laser based systems where a high powered laser pulse is focussed onto a sample surface and induces an ultrasonic pulse which is detected on the opposite sample face by some other type of transducer. Whilst giving high signal amplitudes etc. these systems are expensive, require regular maintenance and are very difficult to use on rough and dirty surfaces. A much cheaper alternative is to use EMATs (Electromagnetic Acoustic Transducers) EMATs have enjoyed comparative success in generating both ultrasonic guided waves and bulk waves in oxides, invars, some steels and particularly aluminium. Their mode of operation means they can be used in the non contact regime and are also much cheaper than laser based systems. However, their low efficiency has precluded their use on many steels due to the lack of appreciable signal amplitude. This is especially true for the case of the spiral coil EMAT, most steels requiring significant signal averaging to discern the ultrasonic echoes induced by the transducer. The work presented in this thesis was aimed at determining which physical properties of the steels had the greatest influence on the electromagnetic transduction mechanism. The mechanism is known to be of purely Lorentz force origin in non magnetic metals whilst magnetostriction is thought to play an important role in ferromagnets. This suggestion has been shown to be incorrect in the case of low tensile carbon and mild steels in that no direct correlation between magnetostriction and spiral coil EMAT performance was found, strongly suggesting a predominantly Lorentzian ultrasonic source. The same was not true for duplex stainless steels and invar that were found to exhibit significantly non linear dependencies of signal amplitude on applied magnetic field. Although it remains unclear whether this was due to a predominantly magnetostrictive source it is almost certainly linked to its two phase microstructure. Measurements of high temperature EMAT efficiency have shown a correlation between the Curie point, Tc, and the disappearance of shear wave signals in all of steels and alloys tested. The emergence of large longitudinal mode signals was also observed at temperatures just above Tc in several of the ferromagnetic samples, although the effect was absent in nickel. Its absence in this particular sample contradicts previously reported observations using the same EMAT geometry by other authors. Use of a modified Michelson interferometer, as a detector, has shown the existence of a frequency doubling effect in all the samples tested, both ferromagnetic and non. The effect was found to be suppressed on application of a magnetic field, this behaviour being consistent with the action of either a 'rectified' magnetostrictive source or a self field Lorentz force. Definitive identification as to which of these is the dominant mechanism was found to be impossible given the small signal amplitudes although the effect itself has not previously been documented as far as the author is aware.