Fast signal predictions of noised signals in eddy current testing

Toshiyuki Takagi, Haoyu Huang, Hiroyuki Fukutomi

Research output: Contribution to journalArticlepeer-review


This paper describes a new method for the simulations of signals noised by the presence of other materials except for test articles in eddy current testing (ECT). In the inspection of steam generator tubes in nuclear power plants, the sensor detects structures, for example support plates to fix the tubes, and deposits adhering to their outside, as well as tube defects. The signals from the structures are. thought to be noises when finding the defects. The method developed here is based on a set of governing equations constructed with magnetic vector potentials, introducing a precondition that a small part of the inverse matrix of the coefficient matrix in an algebraic equation from finite elements and the potentials for the unHawed conductor model are stored. Following the reciprocity theorem, an impedance change due to a flaw, as an eddy current signal, can be computed through integration in the flaw region, and unknowns corresponding to the flaw region only are considered. It results in much smaller degrees of freedom than typical finite element approaches require, and the method provides a fast forward simulator. The fast forward simulator is tested with experimental measurements of ECT benchmark problems with noise factors mentioned above, and is compared with a conventional approach from the point of view of computational costs.

Original languageEnglish
Pages (from-to)2024-2031
Number of pages8
JournalNihon Kikai Gakkai Ronbunshu, A Hen/Transactions of the Japan Society of Mechanical Engineers, Part A
Issue number638
Publication statusPublished - 1999


  • Benchmark problems
  • Deposits
  • Eddy current testing
  • Edge-based finite elements
  • Magnetic vector potentials
  • Reciprocity theorems
  • Steam generator tubes
  • Support plates


Dive into the research topics of 'Fast signal predictions of noised signals in eddy current testing'. Together they form a unique fingerprint.

Cite this