Recognition and Improvement in the Classification of Imperfections Detected by Eddy-Current Inspection
The system presented in this paper shows the interest of image technique for the selectivity improvement of the non destructive benches. It was developed for a Vallourec Group facility that manufactures pipes for the mechanical industry. Installed in the final inspection of the finished products, the system is dedicated to the external soundness inspection by punctual EC. Highly stringent criteria have to be met at this stage to ensure the quality of the end products. Consequently, the rate of false alarms is relatively high in order to keep the client risk near to zero. The purpose of the system is to distinguish the main imperfections that cause these false alarms (e.g. insufficient fusion type flaws), by multi-channels acquisition in real time, followed by 2D imaging at an industrial process rate. The amplitudes of the indications are hen weighted according to the type of imperfection recognized. The system then determines if rejection of the part is founded or if it is a potential false alarm that remains to be verified. The article describes the method used and gives information regarding industrial validation.
Defect classification Punctual EC 2D imagery Tubes
Frēdēric LESAGE Xiaoxing GUO Bernard BISIAUX Jo.l BIONAZ Olivier TAILLEUX Sēbastien PETIT
NDT/Stats Department, CEV-V&M France-Vallourec Group, Aulnoye-Aymeries, 59620 France;
国际会议
第十七届世界无损检测会议(17th World Conference on Nondestructive Testing)
上海
英文
2048-2052
2008-10-25（万方平台首次上网日期，不代表论文的发表时间）