Performance Evaluation of Neural Network Based Pulse-Echo Weld Defect Classifiers

Abstract Pulse-echo ultrasonic signal is used to detect weld defects with high probability. However, utilizing echo signal for defects classification is another issue that has attracted attention of many researchers who have devised algorithms and tested them against their own databases. In this paper, a study is conducted to score the performance of various algorithms against a single echo signal database. Algorithms tested the use of Wavelet Transform (WT), Fast Fourier Transform (FFT) and time domain echo signal features and employed several NN’s architectures such as Multi-Layer Perceptron Neural Network (MLP), Self Organizing Map (SOM) and others known to be good classifiers. The average performance of all can be viewed fair (90%) while some algorithms render success rate of about 94%. It seems that acquiring higher success rates out of a single fixed angle probe pulseecho set up needs new arrangements of data collection, which is under investigation.

[1]  E. Bardal,et al.  Corrosion Detection and Diagnosis , 2007 .

[2]  João M.A. Rebello,et al.  The use of artificial neural network in the classification of pulse-echo and TOFD ultra-sonic signals , 2005 .

[3]  Bachir Boudraa,et al.  Defects Clustering using Kohonen Networks during Ultrasonic Inspection , 2009 .

[4]  Young-Seuk Park,et al.  Self-Organizing Map , 2008 .

[5]  Daniel Massicotte,et al.  Neural classification of Lamb wave ultrasonic weld testing signals using wavelet coefficients , 2001, IEEE Trans. Instrum. Meas..

[6]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[7]  Teuvo Kohonen.  Self-organizing maps of massive document collections , 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium.

[8]  Gene Mathers,et al.  11 – Weld defects and quality control , 2002 .

[9]  Kyungmi Lee,et al.  Classification of Ultrasonic Shaft Inspection Data Using Discrete Wavelet Transform , 2003 .

[10]  David A. Bradley,et al.  The use of neural networks in ultrasonic flaw detection , 1999 .

[11]  S. Mallat.  A wavelet tour of signal processing , 1998 .

[12]  K. Satyanarayana,et al.  Denoising EOG Signal using Stationary Wavelet Transform , 2012 .

[13]  J. A. Ogilvy.  Model for predicting ultrasonic pulse-echo probability of detection , 1993 .

[14]  Redouane Drai,et al.  Time frequency and wavelet transform applied to selected problems in ultrasonics NDE , 2002 .

[15]  Tadeusz Stepinski,et al.  Automatic defect characterisation in ultrasonic NDT , 2000 .

[16]  A. Phinyomark,et al.  Application of Wavelet Analysis in EMG Feature Extraction for Pattern Classification , 2011 .

[17]  J. López-Higuera,et al.  Real-time arc-welding defect detection and classification with principal component analysis and artificial neural networks , 2007 .

[18]  C. Hayes.  The ABC's of nondestructive weld examination , 1997 .

[19]  Jack Blitz.  Fundamentals of Ultrasonics , 1967 .

[20]  R Vicen,et al.  Ultrasonic flaw detection using radial basis function networks (RBFNs). , 2004, Ultrasonics.

[21]  S. F. Burch,et al.  A physical approach to the automated ultrasonic characterization of buried weld defects in ferritic steel , 1986 .