NDT sensor design optimization for accurate crack reconstruction

• Min Li ^[1] ; David Lowther ^[1]
  1. [1] McGill University

     McGill University

     Canadá

     
• Localización: Compel: International journal for computation and mathematics in electrical and electronic engineering, ISSN 0332-1649, Vol. 31, Nº 3 (Special Issue: Problems in Electromagnetism), 2012, págs. 792-802
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • Purpose – The inverse problem related to eddy current testing (ECT) is often formulated as a shape optimization problem. The purpose of this paper is to propose a methodology for determining the optimal parameters of a sensor system for more accurate reconstruction of the crack shape.

    Design/methodology/approach – In this paper, an objective function is formulated using the shape sensitivity information computed from the ECT data. The design of a non‐destructive testing (NDT) sensor is carried out through optimizing the sensor parameters under such a criterion.

    Findings – The methodology proposed results in modifications to the original sensor geometry which makes it more sensitive to the depth changes in a crack. A square wave form of excitation is used in order to provide more information on the size of the crack at different depths, essentially through the superposition of a range of excitation frequencies, each of which has a different depth of penetration. The newly designed ECT sensor system is suitable for dealing with the natural crack problem.

    Research limitations/implications – While the methodology is general and has been shown to work in a simulated environment, the result is not verified by the experiments because the newly designed device has not actually been fabricated.

    Originality/value – This paper has demonstrated the possibility of designing a sensor probe using computer aided design tools without extensive physical testing. The design process is novel and based on a sensitivity approach. This is shown to be very efficient and effective and the solution of the inverse problem demonstrates a very fast convergence.