An EMPECT-based defect localization method for multilayered structures using high-order statistical analysis and high-frequency component separation strategy

The multilayered metallic structures are widely used in industrial fields. For these structures, the internal corrosion defects happening in the process of manufacture and service may seriously affect the availability and safety of the structure and even the whole equipment. In this study, to achieve the effective detection and localization of the internal corrosion defects distributed in different layers, based on energy-management pulsed eddy current testing method, a novel high-order signal feature and a new defect localization method on the basis of high-frequency component separation are proposed. The superiority of the novel high-order signal feature is verified through numerical simulation and experiments. Compared with traditional signal feature, the novel high-order signal feature is able to detect smaller-sized defects and has higher signal-to-noise ratio. The validation of the new defect localization method is proved by experiments. The macro accuracy of the proposed localization method is 91.67 % and this method is still effective for the defect with stepped depth.