Waveform Analysis of Defect Types in Non-destructive Testing


During routine non-destructive testing of ring components, a company observed abnormal waveforms during inspection. The defect-related waveform is shown in Figure 1, which was compared with standard crack defect waveforms commonly encountered in daily inspections (Figure 2). Comparative analysis revealed significant differences: the crack waveform exhibited a single-peak pattern whereas the actual defect displayed a double-peak pattern, indicating substantial variation. To identify the nature of this specific defect type, initial efforts focused on determining its approximate location. Given the small size of the defect, conventional sawing methods might compromise its original morphology and characteristics; thus, wire cutting was employed to ensure precise sampling at the defect site. To establish a correlation between the inspection waveform and the defect features, the sample was dissected to obtain metallographic specimens, followed by low-power acid etching, metallographic examination, and scanning electron microscopy spectroscopy analysis. This study primarily aims to analyze such inspection waveforms to provide a theoretical foundation for subsequent defect identification processes.

40 41

1. Inspection Content

Line-cutting was performed at the approximate location of the defect to obtain low-magnification acid-washed and metallographic samples, which were then subjected to appropriate processing followed by macroscopic examination at low magnification, microscopic analysis, and micro-area evaluation.

1.1 Low-power examination

The defective samples were ground and then subjected to thermal etching in a 1:1 aqueous solution of industrial hydrochloric acid. The microstructure observed under low magnification after acid washing is shown in Figure 3; visual inspection revealed no significant macroscopic segregation defects on the acid-treated surface.

42

1.2 Metallographic Inspection

The defective metallographic samples were ground until the surface was free of scratches and then examined under a metallographic microscope. Analysis of the micrographs revealed that the defects consisted primarily of chain-like inclusions approximately 0.99 mm long (Figure 4), which are aluminum oxide-based chain inclusions. These inclusions severely compromise the mechanical properties of the ring component; during service, they induce stress concentration at the affected location, leading to crack initiation and fracture. Therefore, such inclusions are strictly prohibited. The inclusion classification results are presented in Table 1, with grading criteria compliant with GB/T 10561-2005. The metallographic specimens were etched using a 4% nitric acid-alcohol solution; both the microstructure near the inclusions and the areas without inclusions exhibited identical characteristics—bainite plus fine pearlite—with no coarse-grained structures observed at the defect sites (Figure 5).

43 44 45

1.3 Scanning Electron Microscopy Examination

Microstructural analysis and semi-quantitative elemental composition analysis of the defective metallographic samples using scanning electron microscopy revealed that the inclusions were primarily composed of calcium-aluminum oxides, as shown in Figure 6. These oxides mainly originated from aluminum deoxidizers and calcium treatment agents introduced during the steel refining and deoxidation processes. Therefore, strict control of the refining process is essential to prevent the occurrence of such excessive inclusions, and further quality control measures should be implemented.

Quality inspection and testing frequency are implemented to prevent substandard products exceeding specified limits from leaving the factory.

46

2 Results Discussion

Based on the aforementioned analysis, it is concluded that the occurrence of this type of waveform is attributed to inclusion defects resembling cracks. By comparing these waveforms with those associated with crack defects, distinct waveform spectra corresponding to different defect types can be identified. Examination of the specific flaw detection waveform for this ring component reveals that when the reflection peak exhibits a single, sharp, and leaf-shaped pattern without grass-like variations, the reflected defects primarily consist of cracks, non-metallic inclusions, voids, and fine micro-cracks. Since the reflection interface of non-metallic inclusions is relatively simple, the waveform characteristics lie between those of distinct, sharp, intense peaks and those of dull, slow, flat peaks. A pronounced sharp peak occurs when the interface between the inclusion and metal is smooth or non-viscous; conversely, an attenuated, slow peak appears when the interface is irregularly shaped or highly viscous relative to the metal.

The inspection results indicate that the non-metallic inclusions are calcium-aluminate, classified as endogenous inclusions that are tightly integrated with the metallic matrix. Although their waveforms are relatively sharp and resemble those of microcracks, these inclusions exhibit significant absorption of acoustic energy; consequently, their peak values are lower than those observed in crack waveforms from similar steel grades, and their flaw detection equivalent (Φ2.4) is greater than that of cracks (Φ0.7). The defect area exhibits normal microstructure without coarse grains, hence no grass-like waveforms were detected.

3 Conclusion

(1) Microstructural analysis revealed that the flaw identified by flaw detection is a calcium-aluminate inclusion, classified as an endogenous inclusion, with no coarse crystals observed at the defect site; its flaw detection waveform exhibits a single peak and sharp shape without any grass-like patterns; compared to crack waveforms detected during flaw inspection, this endogenous inclusion defect demonstrates a lower peak amplitude and higher equivalent value.

(2) Defect analysis of this type of waveform provides a basis for distinguishing between crack and inclusion waveforms. This method holds certain practical value for flaw detection inspections.