TL;DR
This paper introduces a novel ultrasonic spiral C-scan technique for non-destructively evaluating stainless-steel resistance spot welds (RSW). By analyzing the geometry of the weld nugget using new "Average of Axis" (AOA) and "Longest to Shortest Axis Ratio" (RLS) factors, the method offers a more efficient and accurate assessment of weld quality and its correlation with mechanical properties, addressing limitations of traditional methods.
What Problem Does This Paper Address?
Modern rail transit relies heavily on stainless-steel bodies, which are joined using resistance spot welding (RSW). The quality of these thousands of spot welds is critical for the safety and stability of the entire vehicle. However, RSW joints are formed in a closed state, making direct observation impossible. Traditional non-destructive testing (NDT) methods like online monitoring of welding parameters, eddy current testing, and conventional ultrasonic C-scans have limitations. Online monitoring is difficult for in-service evaluation, eddy current methods struggle with subsurface defects, and conventional ultrasonic C-scans are often slow and can miss defects in thin materials. Phased array ultrasonic testing, while improving efficiency, can create blind spots. Therefore, there's a clear need for a reliable, efficient, and accurate NDT method to evaluate RSW joint quality, especially for the irregular nugget shapes that can arise.
Methods
The core innovation is the ultrasonic spiral C-scan technique. Unlike traditional X-Y scans that require frequent motor start-stop cycles, leading to mechanical impact and slower speeds, the spiral scan uses a continuous rotational motion. This reduces mechanical stress and significantly increases scanning speed while maintaining resolution.
For evaluating weld quality, the paper introduces two new factors derived from the ultrasonic C-scan images of the nugget zone:
- Average of Axis (AOA) factor: This is calculated by determining the centroid of the irregular nugget morphology and then averaging the lengths of 180 axis lines passing through this centroid at 1-degree intervals. It essentially represents an average measure of the nugget's size.
- Longest to Shortest Axis Ratio (RLS) factor: This factor is derived from the longest and shortest axis lines passing through the centroid. It quantifies the aspect ratio or "roundness" of the nugget.
These factors are proposed as replacements for traditional metrics like nugget diameter and weld aspect ratio (WAR), which can be inaccurate for irregularly shaped nuggets. The study also details the hardware setup, including a portable industrial host and a high-precision rotation detection sensor, and the software flow for data processing.
Datasets
The study utilized 20 joints of double-layer plates welded using a specific resistance spot welding process. These joints were prepared from SUS301L-HT austenitic stainless steel with plate thickness combinations of 4 mm + 4 mm and 2 mm + 2 mm. For evaluating mechanical properties, 16 of these joints were made into tension shear specimens. The paper does not introduce a new benchmark dataset but uses its experimentally prepared weld joints for testing and validation.
Key Insights and Ideas
- Spiral scan efficiency: The spiral C-scan significantly improves detection speed (up to three times faster) and reduces mechanical impact compared to traditional X-Y scans, making it more suitable for inspecting large numbers of welds.
- Beyond simple diameter: Irregular nugget shapes are common in RSW, and traditional metrics like diameter can be misleading. The AOA and RLS factors provide a more robust way to characterize nugget geometry, especially for non-ideal shapes.
- AOA as a primary strength indicator: The AOA factor shows a strong positive correlation with tensile shear strength, suggesting that a larger average nugget size directly translates to a stronger joint. This is intuitive, as a larger area can bear more load.
- RLS as a secondary factor: The RLS factor exhibits a weaker, negative correlation with tensile shear strength. While it influences strength, its effect is less pronounced than AOA, and its precise impact might be complex and depend on orientation.
- Defect detection accuracy: The ultrasonic spiral C-scan method demonstrated good consistency with destructive testing for detecting defect types and sizes, with a detection error of less than 0.1 mm for defect size.
Key Results
- The developed ultrasonic spiral C-scan instrument achieved good consistency in detecting weld nugget morphology and defect types and sizes, visually and accurately displaying them.
- The detection error for defect size was less than 0.1 mm.
- A strong positive correlation (Pearson correlation coefficient of 0.9717) was found between the AOA evaluation factor and the maximum tensile shear force, indicating that as AOA increases, the joint strength increases.
- A negative correlation was observed between the RLS evaluation factor and the strength of the RSW joint, suggesting that higher RLS values (more elongated shapes) generally lead to reduced strength, though this correlation is weaker and more complex.
Limitations and Future Work
- RLS complexity: The paper acknowledges that the correlation between RLS and joint strength is complex and may depend on the orientation of the longest axis relative to the tensile force. Further research and analysis are needed to fully understand this relationship.
- Deep learning integration: Due to the complexity of the weld point strength issue and the large amount of data involved, the authors suggest integrating deep learning techniques in future research plans to further analyze the correlation between RLS and strength.
- Broader application: While the study focuses on stainless steel, further validation on other materials and weld configurations would be beneficial.
Summary updated