Acoustic microscopy signal processing method for detecting near-surface defects in metal materials. (April 2019)

• Record Type: Journal Article
• Authors: Li, Min; Li, Xue; Gao, Chenxing; Song, Yanan
• Is Part Of: NDT & E international. Volume 103(2019)
• Page Start: 130
• Page End: 144
• Publication Date: 2019-04
• Subjects: Ultrasonic detection -- Ultrasonic blind zone -- Mathematical morphology -- Spares deconvolution -- Defect extraction; Nondestructive testing -- Periodicals; Contrôle non destructif -- Périodiques; Electronic journals; 620.1127
• DOI: 10.1016/j.ndteint.2019.02.005
• Languages: English
• ISSNs: 0963-8695

Abstract

Abstract: When ultrasonic pulse echo technique is used for the detection of defects in materials, the amplitude of the defect echo is considerably lower than that of the interface echo, and the echo generated by the defect in the near-surface always overlaps with the interface echo, leading to the difficult extraction of defect characteristics in the near-surface of a material. Therefore, a method combining adaptive morphological filtering with sparse minimum entropy deconvolution (M-S-MED) was proposed. First, adaptive morphological filtering was applied for the removal of background noise and for making the defect echo obvious. Then, sparse minimum entropy deconvolution was performed on the characteristic signals for the acquisition of the reflected pulse sequences of ultrasonic signals. The depth and size of the defect were accurately evaluated because of the effective separation of the interface and defect echoes. The effectiveness of the proposed method was validated by simulating the signals and detecting a near-surface defect in an actual galvanized sheet. The experimental result revealed that the depth and size errors of the actual defect were 1.9% and 3.5%, respectively. Highlights: A new ultrasonic testing method is proposed to characterize near-surface defects. Adaptive morphological filter is used to reduce noise and extract useful features. Non-parametric sparse transformation avoids the influence of human factors. The new method improves the sparsity and temporal resolution of the results. Near-surface defects in galvanized sheets are characterized.