The advent of Giant Magneto-Resistive (GMR) technology permits development of novel highly sensitive array probes for Eddy Current (EC) inspection of multi-layer riveted structures. Multi-frequency GMR measurements with different EC pene-tration depths show promise for detection of bottom layer notches at fastener sites. However, the distortion of the induced magnetic field due to flaws is dominated by the strong fastener signal, which makes defect detection and classification a challenging prob-lem. This issue is more pronounced for ferromagnetic fasteners that concentrate most of the magnetic flux. In the present work, a novel multi-frequency mixing algorithm is proposed to suppress rivet signal response and enhance defect detection capability of the GMR array probe. The algorithm is baseline-free and does not require any assumptions about the sample geometry being inspected. Fastener signal suppression is based upon the random sample consensus (RANSAC) method, which iteratively estimates parameters of a mathematical model from a set of observed data with outliers. Bottom layer defects at fastener site are simulated as EDM notches of different length. Performance of the proposed multi-frequency mixing approach is evaluated on finite element data and experimental GMR measurements obtained with unidirectional planar current excitation. Initial results are promising demonstrating the feasibility of the approach.The advent of Giant Magneto-Resistive (GMR) technology permits development of novel highly sensitive array probes for Eddy Current (EC) inspection of multi-layer riveted structures. Multi-frequency GMR measurements with different EC pene-tration depths show promise for detection of bottom layer notches at fastener sites. However, the distortion of the induced magnetic field due to flaws is dominated by the strong fastener signal, which makes defect detection and classification a challenging prob-lem. This issue is more pronounced for ferromagnetic fasteners that concentrate most of the magnetic flux. In the present work, a novel multi-frequency mixing algorithm is proposed to suppress rivet signal response and enhance defect detection capability of the GMR array probe. The algorithm is baseline-free and does not require any assumptions about the sample geometry being inspected. Fastener signal suppression is based upon the random sample consensus (RANSAC) method, which iteratively estimates parameters of ...
[1]
Robert C. Bolles,et al.
Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography
,
1981,
CACM.
[2]
Lalita Udpa,et al.
Analysis of EC-GMR data for detection of cracks under fasteners (CUFs)
,
2016
.
[3]
Guang Yang,et al.
EC-GMR array with rotating current excitation for multilayered riveted structures inspection
,
2015
.
[4]
Lalita Udpa,et al.
Image processing algorithms for automated analysis of GMR data from inspection of multilayer structures
,
2015
.
[5]
Lalita Udpa,et al.
Sensor-tilt invariance analysis for eddy current signals
,
2012
.
[6]
Lalita Udpa,et al.
3D EC-GMR sensor system for detection of subsurface defects at steel fastener sites
,
2012
.
[7]
L. Udpa,et al.
Reduced Magnetic Vector Potential Formulation in the Finite Element Analysis of Eddy Current Nondestructive Testing
,
2009,
IEEE Transactions on Magnetics.
[8]
John C. Aldrin,et al.
Enhanced image processing methods for GMR array inspections of multilayer metallic structures
,
2013
.