Optimizing Steel Coil Production: An Enhanced Inspection System Based on Anomaly Detection Techniques

Endless strip generation is the key to productivity and quality in several types of steel coil production lines. Coil-tocoil joining by means of welding machines provides such a strip. Since the joint is the weakest area of the strip, its quality must be assessed for the production line to accept it. Therefore, it is necessary to inspect the quality of the weld in the welding-cycle time. Based on our knowledge acquired in the previous development of quality assessment prototypes for steel strips, we present in this article an improved inspection system to detect defective resistance seam welds based on anomaly detection techniques. This system does not rely on the weld classifications done in production lines based on welding control programs. Therefore, it is immune to the influence of both incorrectly configured welding control programs and chemical composition variations from coil to coil of the same steel grade. Tuning the inspection system required a fully experimental design, which would have taken several months using a conventional computer. For this reason, the high-performance computing (HPC) facilities at the Edinburgh Parallel Computing Center were used to cut down the tuning time.

[1]  T. Warren Liao,et al.  Improving the accuracy of computer-aided radiographic weld inspection by feature selection , 2009 .

[2]  Hao Wang,et al.  Detection of line weld defects based on multiple thresholds and support vector machine , 2008 .

[3]  Janez Diaci,et al.  Estimating the strength of resistance spot welds based on sonic emission , 2005 .

[4]  Juan Zapata-Pérez,et al.  Performance evaluation of an automatic inspection system of weld defects in radiographic images based on neuro-classifiers , 2011, Expert Syst. Appl..

[5]  Wei Zhang,et al.  Application of infrared imaging for quality inspection in resistance spot welds , 2009, Defense + Commercial Sensing.

[6]  D.F. Garcia,et al.  Real-Time Assessment of the Reliability of Welds in Steel Strips , 2010, IEEE Transactions on Industry Applications.

[7]  Roman Gr. Maev,et al.  Ultrasonic quality control methods for spot-welded joints , 2011 .

[8]  Sehun Rhee,et al.  Quality estimation of resistance spot welding by using pattern recognition with neural networks , 2004, IEEE Transactions on Instrumentation and Measurement.

[9]  J. Molleda,et al.  Fuzzy-based approach to real-time detection of steel strips defective welds , 2005, CIMSA. 2005 IEEE International Conference on Computational Intelligence for Measurement Systems and Applications, 2005..

[10]  Yue Min Zhu,et al.  Uncertainty modelling using Dempster-Shafer theory for improving detection of weld defects , 2003, Pattern Recognit. Lett..

[11]  Daniel F. García,et al.  A fast and robust decision support system for in-line quality assessment of resistance seam welds in the steelmaking industry , 2012, Comput. Ind..

[12]  Y. Zhou,et al.  Dynamic Electrode Force and Displacement in Resistance Spot Welding of Aluminum , 2004 .

[13]  Du Dong Image processing algorithm for weld seam recognition based on color analyzing , 2009 .

[14]  Şükrü Talaş,et al.  The assessment of carbon equivalent formulas in predicting the properties of steel weld metals , 2010 .