Ebook: Electromagnetic Nondestructive Evaluation (XIV)

This volume contains selected papers from the fifteenth International Workshop on Electromagnetic Nondestructive Evaluation, which was held in Szczecin, Poland, from June 13 to 16, 2010.

Previous ENDE Workshops have been held in: London, United Kingdom (1995); Tokyo, Japan (1996); Reggio Calabria, Italy (1997); Chatou, France (1998); Des Moines, United States (1999); Budapest, Hungary (2000); Kobe, Japan (2001); Saarbrücken, Germany (2002); Paris, France (2003); East Lansing, United States (2004); Iwate, Japan (2006); Cardiff, United Kingdom (2007); Seoul, Korea (2008); Dayton, United States (2009).

The aim of the workshop, organized by the West Pomeranian University of Technology, Szczecin, Poland and the Japanese Society of Maintenology, was to bring together scientists from universities and research institutions conducting in-depth research into the basics of electromagnetic non-destructive evaluation (ENDE) on the one hand, and engineers presenting practical problems and industrial applications on the other.

Ninety nine participants from eleven European countries and from Algeria, Australia, Brazil, China, India, Japan, Korea, and the United States, were officially registered. Eighty papers were presented in all, among them five invited papers, namely:

1. D.C. Jiles, Ł.P. Mierczak, Y. Melikhov, Detection of Surface Condition in Ground Steel Components Using Magnetic Barkhausen Measurements,

2. S. Honda, Pipe Wall Thickness Inspection with Current Driven Thermal Method,

3. J.S. Knopp, M. Blodgett, J. Calzada, E. Lindgren, C. Buynak, J. Aldrin, Computational Methods in ENDE: Revolutionary Capability for the Sustainment of Aerospace Systems in the 21st Century,

4. Z. Chen, T. Hwang, L. Wang, S. Tian, N. Yusa, Investigation on the Features of the Electric Conductivity Around a Stress Corrosion Crack,

5. S.C. Chan, R. Grimberg, J.A. Hejase, Z. Zeng, P. Lekeakatakunju, L. Udpa, S.S. Udpa, Development of a Nonlinear Eddy Current Technique for Estimating Case Hardening Depths.

Short versions of all the contributions have been published in the Book of Abstracts, and reviewed and accordingly revised full papers have been accepted and are now included in this volume: Electromagnetic Non-Destructive Evaluation (XIV) published by IOS Press in the series Studies in Applied Electromagnetics and Mechanics.

In closing, we would like to thank the authors, session chairs, and reviewers for conscientiously executing their duties to maintain the high scientific quality of the papers published in this volume. We believe that the readership of this book will find the included papers interesting and inspiring.

T. Chady, S. Gratkowski, T. Takagi, S.S. Udpa

Co-Editors

Electric potential difference(EPD) method is widely used for diagnosing and monitoring flow conduits. Theoretical analysis and the exact solution of the electrical potential field in pipe wall have been given, and 2D approximation is effective for small pipe wall thickness there. This study proposes to use temperature distribution driven by Joule heat with the electric current other than electrical potential difference. Under thin wall approximation, 2D electric and thermal problems are analyzed and exact solutions are given.

In any NDE application, the availability of a theoretical model is very useful in understanding the capability of the system since it can enable the visualization of field/flaw interaction for different flaw and sensor geometries. A numerical model capable of simulating inspection of multi-layered structures using eddy current excitation and magneto-resistive sensing is developed to predict 2D image data of rivet locations. The model is used to study particularly the effect of sensor tilt on the probability of detection (POD) of defects at rivet sites. The modeling of sensor-tilt during inspection around rivet sites has to take into account finite source coil and meshing of the finite source coil at each scan position. This paper presents a finite element (FE) model formulation using reduced magnetic vector potential and electric scalar potential in a way that does not require meshing source coils. In addition to simulating the effect of sensor tilt on the signal, this paper also describes features in the signal that are invariant to sensor tilt but sensitive to defect parameters.

A non-iterative MUSIC-type imaging algorithm for time-harmonic eddy-current testing of a damaged conductive plate is considered. The method is based (i) on a first-order asymptotic formulation of the secondary magnetic field observed outside the plate and induced by volumetric defects within it, which is valid when those are small enough with respect to the skin-depth at the frequency of operation, and (ii) on the singular value decomposition of the multistatic response matrix that can be collected using an appropriate source and sensor array. Herein, the asymptotic field formulation is validated by comparing the results with those provided by means of the CIVA platform in a realistic case. Further, the singular value pattern and MUSIC images are investigated, for a small spherical void placed at different locations inside a plate. A Gaussian distributed noise is added to the asymptotic data during the initial investigation, to alleviate inverse crime at least in part (more complex arrangements of source/sensor arrays and defects will be considered with external data in a later stage). Yet, these preliminary results as exhibited indicate the potential of the method to locate small defects in conductive materials in a typical eddy-current testing configuration

A radiographic system for identification of defects in three-dimensional space has been built and tested. It is a cost effective alternative to sophisticated fully three-dimensional tomography systems. We present initial results of using the extended radiographic system for the images of the phantom made of wood and pieces of copper.

Eddy-current induction by a coaxial circular air-cored coil in a cracked borehole is considered in the limit of small electromagnetic skin depth. A closed form expression is derived for the change in coil impedance ΔZ due to a long axial crack with constant depth and opening based on an extension of the Lewis extended surface impedance boundary conditions to a cylindrical geometry. The theoretical predictions are compared with experimental measurement of ΔZ for a series of boreholes containing wire-cut slots in 2024 Al alloy test specimens. The model predictions were in good agreement with experiment for smaller slot depths, with the accuracy tending to decrease as the slot depth increased. In all cases, the predictions were superior to those based on a simple Born approximation.

The determination of magnetic distortion fields caused by inclusions hidden in a conductive matrix using homogeneous current flow needs to be addressed in multiple tasks of electromagnetic non-destructive testing and materials science. This includes a series of testing problems such as the detection of tantalum inclusions hidden in niobium plates, metal inclusion in a nonmetallic base material or porosity in aluminum laser welds. Unfortunately, straightforward tools for an estimation of the defect response fields above the sample using pertinent detection concepts are still missing. In this study the Finite Element Method (FEM) was used for modeling spherically shaped defects and an analytical expression developed for the strength of the response field including the conductivity of the defect and matrix, the sensor-to-inclusion separation and the defect size. Finally, the results also can be useful for Eddy Current Testing problems, by taking the skin effect into consideration.

The paper deals with uncertainty in depth estimation of partially conductive cracks from eddy current testing signals. Numerical simulations are carried out to investigate influences of crack parameters on the response signals. Three parameters of the crack, i.e. its depth, width and conductivity are changed to simulate variety of possible scenarios. Standard pancake probe is used for the inspection at first. The presented numerical results confirm that the inverse problem is highly ill-posed when partially conductive cracks, as stress corrosion cracking, are concerned. Accordingly, a new approach in the response signal sensing is proposed. The idea is to pick-up more information provided by the eddy currents. Thus, three coils positioned perpendicularly to each other are employed to sense all the three spatial components of the perturbation electromagnetic field. The gained results reveal that it is possible to considerably decrease the uncertainty in depth estimation of partially conductive cracks by employing the proposed idea.

Semi-discrete method is known since 80's of XX^th century. The method provides analytical solution in time, so the time-stepping may be omitted. Comparing to usual finite elements method in time, this method seems not to be numerically effective, because produced matrices are dense. Because of this reason, it was rather rarely used. However to carry out sensitivity analysis with adjoint models [1] solutions in forward and backward time have to be obtained. Both time points should coincide. For space discretization finite elements are used, as usual. The semi-discrete method allows to determine analytically the continuous solution for any given time of analysis. In this paper evaluation of this method for different kinds of excitation shapes is presented. The sensitivity analysis is applied for inverse task of recognition of conductivity distribution in non-destructive testing of materials using eddy-currents.

A method to characterize defects by processing eddy-current testing (ECT) signals is presented. It works in a short time for generic defects and work-pieces. To retrieve the dimensions of a defect, a metamodel-based Particle Swarm Optimization (PSO) is considered. Indeed, using a metamodel as an input of a stochastic method enables to significantly speed it up. The metamodel is generated using an adaptive database generation. Both tube and plate configurations, corresponding with either simulated or laboratory-controlled measured data are considered as illustration. Good accuracy and satisfactory speed of the method are exhibited, additional information besides the inversion results being provided as well, which highlights possibly indeterminate cases, making the method useful also for decision analysis.

The motivation of this work is to develop a computational treatment of eddy current simulation applicable to complex 3D geometries. To this end, an approach based on surface integral equations and involving the homogeneous full-space Green's function and dyad is developed. A Galerkin variant of the method of moments, using bilinear basis functions, is employed for discretizing the surface integral equations, leading to a linear system. Solving the linear system using a direct matrix solver is prohibitively expensive in computing time and memory resources. This difficulty is circumvented by using an iterative solver (typically GMRES) together with the Fast Multipole Method (FMM). The response of an eddy current probe is finally expressed with the Auld reciprocity theorem.

This paper deals with the development of a fast numerical semi-analytical model dedicated to the simulation of the non-destructive testing of ferromagnetic tubes by using the magnetic flux leakage method(MFL). Taking into account the characteristic of the ferromagnetic material, the B-H curve, the numerical model is based on the generalized boundary element method which implies the evaluation of two magnetic fictitious scalar densities. Integral equations are discretized via the Galerkin's method and the particularity of this paper lies in the implementation of high order basis functions for the interpolation of the scalar unknowns. Simulated data provided by the numerical model are compared to FE data in order to show the validity of the numerical approach.

In simulations of eddy current inspections using the boundary element method (BEM) is is often the case that the calculation of the system matrix carries the largest computational cost. This matrix calculation typically makes use of a Green's function expressed as a Fourier integral. The present study has aimed to reduce the computational cost by adopting an alternative approach for the rapid computation of the Green's function in the spatial domain. This is possible by employing Sommerfeld integral expressions and by approximating them in terms of discrete complex images. Theoretical results are compared to both benchmark experimental data for a semi-elliptical crack and to numerical data from a finite element commercial package. Agreement is excellent over a wide frequency range.

In this work for the first time we validate experimentally a fast (real-time) non-iterative imaging method for eddy current tomography of conductive materials. The presented method is based on the monotonicity principle applied in the low frequency limit where the skin-depth is great with respect to the relevant dimension of the problem under test. The experimental tests prove the effectiveness of the monotonicity based imaging method. Moreover, we prove that the inversion can be carried out without resorting the direct problem solver.

This study evaluates the applicability of using microwaves for the inspection of pipe wall thinning by using numerical simulations and experiments. The configuration considered in this study is an axisymmetric one consisting of a straight brass tube and coaxial cables. After the agreement between three-dimensional numerical simulations and experimental result is confirmed, experiments to locate wall thinning are carried out. Clear signals due to the wall thinning are obtained. The time of flight of microwaves agrees with the theoretical prediction, which demonstrates the possibility of locating wall thinning from microwave signals.

A calculation method for the simulation of the Eddy Current Testing (ECT) signal of multiple cracks is implemented on a general purpose video card. NVIDIA's parallel computing solution, the Compute Unified Device Architecture (CUDA) was used as a shell for the implementation on both hardware and software sides. Possibilities inherent in this technology are shown through an example and a detailed implementation focusing on data parallel execution is discussed. Different realizations of the same algorithm are inspected and a comparison from the view of efficiency is given.

In contemporary non-destructive inspections Electromagnetic-Acoustic Transducers (EMATs) become more and more attractive alternative in comparison to conventional contact piezoelectric sensors. Such a technique does not require a mechanical coupling and inspection can be performed in the presence of an insulation or coating. Another advantage is the possibility of inspecting in higher temperature environments. In this paper we discuss different aspects of chosen configurations of Electromagnetic-Acoustic Transducers. Numerical modeling was carried out for two types of shear horizontal wave EMATs, i.e. Periodic Permanent Magnets EMAT and Meander-Line Coil EMAT, placed on conducting and nonmagnetic plate. Preliminary signal analysis based on 2D FFT has been introduced.

Eddy Current Testing (ECT) is a standard technique in industry for the detection of surface breaking flaws in ferromagnetic materials such as steels. In this context, simulation tools can be used to improve the understanding of experimental signals, optimize the design of sensors or evaluate the performance of ECT procedures. CEA has developed for many years semi-analytical models embedded into the simulation platform CIVA [1] dedicated to non-destructive testing. Following a previous work carried out at the laboratory in the case of one flaw located in a cylindrical ferromagnetic part, the developments presented herein address the case of a flaw located inside a planar and ferromagnetic medium. The theoretical approach will be presented, as well as comparisons between simulation data obtained from the literature.

Eddy current models have matured to such a degree that it is now possible to simulate realistic nondestructive inspection (NDI) scenarios. Models have been used in the design and analysis of NDI systems and to a limited extent, model-based inverse methods for Nondestructive Evaluation (NDE). The science base is also being established to quantify the reliability of systems via Model-Assisted Probability of Detection (MAPOD). In realistic situations, it is more accurate to treat the input model variables as random variables rather than deterministic quantities. Typically a Monte-Carlo simulation is conducted to predict the output of a model when the inputs are random variables. This is a reasonable approach as long as computational time is not too long; however, in most applications, introducing a flaw into the model results in extensive computational time ranging from hours to days, prohibiting Monte-Carlo simulations. Even methods such as Latin-Hypercube sampling do not reduce the number of simulations enough for reasonable use. This paper presents the Probabilistic Collocation Method as a non-intrusive alternative to other uncertainty propagation techniques.

Eddy current techniques are useful for the detection and the characterization of flaws in conductive structures. The flaw signal results from the measurement of the local punctual value of the perturbed magnetic field above the metal surface in the vicinity of the defect by using some magnetic field sensor (Giant Magneto Resistance sensor (GMR) or Giant Magnetic Impedance sensor (GMI). This paper aims to describe a semi-analytical numerical model integrated into the CIVA platform which is able to compute the flaw signal due to a given notch embedded in a planar stratified media. The inducer may be chosen in a list of available exciting coils into CIVA. Some numerical experiments and a comparison between simulated data and experimental data show the validity of the new developments.

In this paper the authors focus on detection of wire image quality indicators, which are widely applied in industrial radiography. Two algorithms for IQI detection are proposed. The first algorithm is intended for fast and rough detection of position of IQIs. It is based on evaluation of correlation between the radiographic image of IQI (together with a plastic case) and the fragments of the analyzed radiogram. The second algorithm uses the Radon transform applied to the region of interest (ROI) obtained from the first algorithm to extract the exact position of wires as well as information on visibility of each wire.

In this work, a code of finite element method is developed for numerical simulation of the acoustoelastic effect in pre-stressed media. The possibility of assessing the stress status, by using the EMAT (electromagnetic acoustic transducer) receiver for precise measurement of LCR (refracted longitudinal) wave, is investigated. Besides that, the relationship between the polarization of transient Rayleigh wave and the state of stress is also predicted by simulation.

BEMLAB [1] is the open source project implementing Boundary Element Method (BEM) [2,3] comprehensively. BEMLAB binary packages and the source code are distributed under GNU LGPL (Lesser General Public License) license terms. The project provides universal library and the reference application, which is the easiest way for solving problems using BEM. The article presents example BEMLAB application in modelling comb capacitors used in Micro-Electro-Mechanical Systems (MEMS). Finally capacitance will be calculated using BEM results. Particularly Asymptotic Boundary Conditions (ABC) will be used to model external space of the capacitor (the geometry of infinite half-lane) as an internal BEM problem. This approach will allow to calculate total capacitance with dispersed capacitance included. Results will be compared with the simplified model.