Automatic detecting and classifying defects during eddy current inspection of riveted lap-joints
Uppsala University, Signals and Systems, SE-751 20 Uppsala, Sweden NDT & E International
(Impact Factor: 2.23).
01/2000; 33(1):47-55. DOI: 10.1016/S0963-8695(99)00007-9
This article presents a novel method for automatic detection and classification of cracks located in the second lower layer of the aircraft lap-joints during Eddy Current (EC) inspection. The cracks originating from the rivet holes were detected using a tailor-made deep penetrating EC probe. The proposed method consists of three steps: pre-processing, feature extraction and classification. The pre-processing, performed before the feature extraction included median filtering, rotation and de-biasing of the EC patterns. The rotation of the patterns was performed so that energy of the responses to the rivets was maximized along the quadrature direction, while the defect responses were maximized in the in-phase direction in the impedance plane. Feature extraction was then performed using four different methods: discrete wavelet transform, Fourier descriptors, principal component analysis (PCA) and block mean values. The classification was performed using a standard multi-layer perceptron (MLP) neural network. All the pre-processing methods showed similar classification performance on the used data set, but the PCA method compressed the data best.
Available from: Luis S. Rosado
- "However , this approach requires nonlinear-regression algorithms to approximate the data introducing a significant complexity. The study of a neural network classifier using features extracted with several methods was reported in . The discrete wavelet transform, Fourier descriptors, principal component analysis, and block mean values to extract features from ECT data were considered. "
[Show abstract] [Hide abstract]
ABSTRACT: The estimation of the parameters of defects from eddy current nondestructive testing data is an important tool to evaluate the structural integrity of critical metallic parts. In recent years, several works have reported the use of artificial neural networks (ANNs) to deal with the complex relation between the testing data and the defect properties. To extract relevant features used by the ANN, principal component analysis, wavelet decomposition, and the discrete Fourier transform have been proposed. In this paper, a method to estimate dimensional parameters from eddy current testing data is reported. Feature extraction is based on the modeling of the testing data by a template of additive Gaussian functions and nonlinear regressions to estimate their parameters. An ANN was trained using features extracted from a synthetic data set obtained with finite-element modeling of the eddy current probe. The proposed method was applied to both simulated and measured data, providing good estimates.
Available from: Irene Epifanio
- "Other interesting applications include medical diagnosis from EEG measurements from multiple scalp sites (Anderson et al., 1998), the automatic classification of rivet defects using eddy currents (Lingvall and Stepinski, 2000), gait analysis (Huang, 2001) and chemometric applications such as the prediction of the fat content of a meat sample based on the nearinfrared absorbance spectrum (Ferraty and Vieu, 2003) or a polymer discrimination problem (Naya et al., 2004). "
[Show abstract] [Hide abstract]
ABSTRACT: Curve discrimination is an important task in engineering and other sciences. We propose several shape descriptors for classifying functional data, inspired by form anal-ysis from the image analysis field: statistical moments, coefficients of the components of independent component analysis (ICA) and two mathematical morphology descrip-tors (morphological covariance and spatial size distributions). They are applied to three problems: an artificial problem, a speech recognition problem and a biomechan-ical application. Shape descriptors are compared with other methods in the literature, obtaining better or similar performances.
Available from: Radislav Smid
- "However, modelling based analysis methods is difficult to apply for test data evaluation, so in automated analysis systems the EC signatures as a trajectory of the signal in the impedance plane are usually described using simple physical features such as maximum amplitude, amplitude projections in the complex plane, various phase angles and phase spread , . This approach is often not sufficiently robust and universal, and other researchers have investigated more sophisticated features such as coefficients of the Fourier transform , the discrete wavelet transform (DWT) of the in-phase signature component , principal component analysis (PCA) , and angle signature . "
[Show abstract] [Hide abstract]
ABSTRACT: This article presents a novel method for automatic evaluation of flaws during a manual eddy current (EC) inspection procedure. In manual scanning, a signal related to impedance change in the complex plane is affected by large and unpredictable variations of scanning speed and alterations of probe position. This paper introduces a robust EC signal normalization method using non-linear filtration based on evaluation of the distance between consecutive samples in the complex plane and median based tracking of the EC signature. Feature extraction was performed using normalized Fourier and complex discrete wavelet descriptors. The classification was performed using six different methods: nearest-mean classifier, k-nearest neighborhood classifier, the standard multi-layer feed forward neural network with backpropagation, radial basis network, support vector machines and the adaptive-network-based fuzzy inference system. The method was tested on a single frequency instrument with an absolute probe and a dual frequency instrument with a probe for fast testing of rivets in layered structures. The results obtained using this approach demonstrate the effectiveness of the proposed system.
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.