Pipelines are the safest and most economical method of transporting oil over long distances. However, the risk of leakage caused by discontinuities in the pipeline wall is always present. The most commonly used methods for the nondestructive testing (NDT) of riser welds are radiography and ultrasound. The use of the ultrasonic process is increasing due to automation of the technique and because of the difficulties found in radiographic testing to detect cracks that are parallel to the X-ray beam (difficulties which have been eliminated by using the ultrasonic time of flight diffraction technique), as well as the high costs involved in radiographic testing. Although the time of flight diffraction technique can reach high speeds and have a high detection probability and low rate of incorrect results, the classification of discontinuities through this technique is still frequently questioned, since it depends mainly on the expertise of the operator. The purpose of this study is to test pattern recognition by neural networks and signal processing in the classification of welding discontinuities in rigid risers made of low carbon steel with a nominal diameter of 152 mm (6 in.) and a thickness of 11 mm (0.4 in.), when referring to these three conditions of the weld joint: no discontinuity; lack of fusion; and lack of penetration. Backpropagation networks were efficient in classifying these kinds of signals.
Automatic testing of rigid risers using time of flight diffraction and neural networks
Automatische Prüfung von starren Steigleitungen mit Laufzeitbeugung und neuronalen Netzen
Materials Evaluation ; 62 , 10 ; 1066-1070
2004
5 Seiten, 8 Bilder, 4 Tabellen, 16 Quellen
Aufsatz (Zeitschrift)
Englisch
Comparative Analysis of Design Criteria for Rigid Risers
British Library Conference Proceedings | 2007
|Dynamic compression of rigid and flexible risers: Experimental and numerical results
Tema Archiv | 2006
|