A New System for Steam Boiler Tubes using Artificial Neural Network

Abstract

A major inspection challenge facing the boiler industries is to satisfy the welds in tubing and plate during the manufacturing, erection and commissioning stages. During commercial operation the plant boiler tubes that leaks due to downtime. As the main part of the boiler, the water wall tube which runs on dangerous environment may lead to serious boiler accidents, causing of the tubes failure. For this reason, inspecting the defects on the steam boiler tube comprehensively is extremely required. However, single traditional NDT inspection method is limited .In terms of the ability of identifying different kinds of defects. In order to meet the requirement of fullscale inspection, in this paper, a new system foe steam boiler tubes are experimented. After experiment, the new system can effectively measure the remaining wall thickness and identify different kinds of defects including pinholes and circumferential cracks with the sensor being moved outside the tubes using artificial neural networks.