Signal Processing Schemes for Defect Identification and Sizing From Signal of Eddy Current Testing in Nuclear Power Plant

Abstract

In this paper, signal processing schemes developed for upgrading the Eddy Current Testing (ECT) system utilized for In Service Inspection (ISI) of Steam Generator (SG) tubes of a nuclear power plant are presented. Schemes for the wobbling noise recognition, signal indication extraction, signal classification, and the separation of mixed signals are proposed. The schemes are applied to both signals measured in a SG mock-up and signals from short tube test-pieces measured in a laboratory environment. It is demonstrated that the methods work well even for mixed signals in the support plate region of SG tubes. As it is possible to adjust the position of a signal point based on the signature of the support plates, noise signals from a welding line or a bending zone can be recognized based on the information of design and manufactory. The mixed signals of defects and support structures are processed by using a similarity analysis strategy and the identification of defect signals is performed by means of both a neural network approach and a statistic method. Satisfactory results are obtained for the processing of measured signals.