Real-Time Testing of Steel Strip Welds based on Bayesian Decision Theory

One of the main trouble in a steel strip manufacturing line is the breakage of whatever weld carried out between steel coils, that are used to produce the continuous strip to be processed. A weld breakage results in a several hours stop of the manufacturing line. In this process the damages caused by the breakage must be repaired. After the reparation and in order to go on with the production it will be necessary a restarting process of the line. For minimizing this problem, a human operator must inspect visually and manually each weld in order to avoid its breakage during the manufacturing process. The work presented in this paper is based on the Bayesian decision theory and it presents an approach to detect, on real−time, steel strip defective welds. This approach is based on quantifying the tradeoffs between various classification decisions using probability and the costs that accompany such decisions. Keywords—Classification, Pattern Recognition, Probabilistic Reasoning, Statistical Data Analysis.

[1]  Karl Deutsch.  Automated Ultrasonic Inspection , 2005 .

[2]  Madjid Fathi,et al.  Vision systems for the inspection of resistance welding joints , 2000, Electronic Imaging.

[3]  Yagmur Denizhan,et al.  Classification trees prove useful in nondestructive testing of spot weld quality , 1993 .

[4]  Alvin M. Strauss,et al.  Automated visual inspection and interpretation system for weld quality evaluation , 1995, IAS '95. Conference Record of the 1995 IEEE Industry Applications Conference Thirtieth IAS Annual Meeting.

[5]  Xiao-Guang Zhang,et al.  Defects recognition on X-ray images for weld inspection using SVM , 2004, Proceedings of 2004 International Conference on Machine Learning and Cybernetics (IEEE Cat. No.04EX826).

[6]  Leopoldo García Franquelo,et al.  Neural network approach to weld quality monitoring , 1994, Proceedings of IECON'94 - 20th Annual Conference of IEEE Industrial Electronics.

[7]  John W. Sheppard,et al.  A Bayesian approach to diagnosis and prognosis using built-in test , 2005, IEEE Transactions on Instrumentation and Measurement.

[8]  Sang-Ryong Lee,et al.  Neuro-fuzzy algorithm for quality assurance of resistance spot welding , 2000, Conference Record of the 2000 IEEE Industry Applications Conference. Thirty-Fifth IAS Annual Meeting and World Conference on Industrial Applications of Electrical Energy (Cat. No.00CH37129).

[9]  Yi Sun,et al.  Real-time weld defect inspection system in x-ray images , 2002, SPIE/COS Photonics Asia.

[10]  T. Zacharia,et al.  Neural network-based resistance spot welding control and quality prediction , 1999, Proceedings of the Second International Conference on Intelligent Processing and Manufacturing of Materials. IPMM'99 (Cat. No.99EX296).

[11]  Zeng An,et al.  Welding quality monitoring and management system based on data mining technology , 2003, Proceedings of the 2003 International Conference on Machine Learning and Cybernetics (IEEE Cat. No.03EX693).

[12]  Ingvar Claesson,et al.  Quality monitoring in robotised welding using sequential probability ratio test , 1996, Proceedings of Digital Processing Applications (TENCON '96).

[13]  Shigeo Abe DrEng.  Pattern Classification , 2001, Springer London.

[14]  David G. Stork,et al.  Pattern Classification , 1973 .

[15]  Stephen W. Kercel,et al.  In-process detection of weld defects using laser-based ultrasound , 1999, Optics East.