Inverse mapping of magnetic flux leakage signal for defect characterization

Abstract Estimation of shape and size of metal loss defects in ferromagnetic plates/pipes from magnetic flux leakage (MFL) signals calls for the development of an inverse mapping system, which is very often non-linear and ill-posed. Identification of such systems needs regularization and modeling in multi-resolution framework. This paper proposes a new inverse mapping approach for defect profile estimation and sizing from MFL signals, using a partially time-varying model. The model works with wavelet projections and selectively employs time varying and time invariant models in different sub-bands to estimate the defect shape. Shape synthesis uses alternate projections (in two signal spaces) that converge to minimum norm solution. Length and width profile of defects are estimated independently. The model is trained with reference defects and is validated with unknown signals from laboratory and field runs in test rigs. This paper reports 1-D results which show excellent match with actual sizes and shapes of defects.

[1]  Stéphane Mallat,et al.  A Theory for Multiresolution Signal Decomposition: The Wavelet Representation , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  G. P. Srivastava,et al.  Empirical structure for characterizing metal loss defects from radial magnetic flux leakage signal , 2010 .

[3]  D Marr,et al.  Theory of edge detection , 1979, Proceedings of the Royal Society of London. Series B. Biological Sciences.

[4]  John S. Baras,et al.  Properties of the multiscale maxima and zero-crossings representations , 1993, IEEE Trans. Signal Process..

[5]  Stephen A. Billings,et al.  A new class of wavelet networks for nonlinear system identification , 2005, IEEE Transactions on Neural Networks.

[6]  Jay A. Farrell,et al.  Wavelet-based system identification for nonlinear control , 1999, IEEE Trans. Autom. Control..

[7]  Marco Gori,et al.  On the problem of local minima in recurrent neural networks , 1994, IEEE Trans. Neural Networks.

[8]  R. Ghanem,et al.  A WAVELET-BASED APPROACH FOR THE IDENTIFICATION OF LINEAR TIME-VARYING DYNAMICAL SYSTEMS , 2000 .

[9]  Ikhlas Abdel-Qader,et al.  ANALYSIS OF EDGE-DETECTION TECHNIQUES FOR CRACK IDENTIFICATION IN BRIDGES , 2003 .

[10]  Jozef Vörös,et al.  Parameter identification of Wiener systems with multisegment piecewise-linear nonlinearities , 2007, Syst. Control. Lett..

[11]  Martin Vetterli,et al.  Discrete-time wavelet extrema representation: design and consistent reconstruction , 1995, IEEE Trans. Signal Process..

[12]  Roberto Kawakami Harrop Galvão,et al.  Wavelet-packet identification of dynamic systems in frequency subbands , 2006, Signal Process..

[13]  Stéphane Mallat,et al.  Characterization of Signals from Multiscale Edges , 2011, IEEE Trans. Pattern Anal. Mach. Intell..

[14]  Stéphane Mallat,et al.  Zero-crossings of a wavelet transform , 1991, IEEE Trans. Inf. Theory.

[15]  Lennart Ljung,et al.  System identification (2nd ed.): theory for the user , 1999 .

[16]  Tao Jin,et al.  Designing and signal processing of intelligent inspection pig applying magnetic flux leakage methods , 2004, 2004 International Conference on Intelligent Mechatronics and Automation, 2004. Proceedings..

[17]  Marco S. Reis,et al.  A multiscale empirical modeling framework for system identification , 2009 .

[18]  Gui Yun Tian,et al.  3D magnetic field sensing for magnetic flux leakage defect characterisation , 2006 .

[19]  Vikram M. Gadre,et al.  Function learning using wavelet neural networks , 2000, Proceedings of IEEE International Conference on Industrial Technology 2000 (IEEE Cat. No.00TH8482).

[20]  G. P. Srivastava,et al.  Characterisation of metal loss defects from magnetic flux leakage signals with discrete wavelet transform , 2000 .

[21]  John Mason,et al.  Experimental study to differentiate between top and bottom defects for MFL tank floor inspections , 2009 .