In situ detection of welding defects: a review

• Published in: Welding in the world Vol. 66; no. 4; pp. 611 - 628
• Main Authors: Madhvacharyula, Anirudh Sampath, Pavan, Araveeti V Sai, Gorthi, Subrahmanyam, Chitral, Srihari, Venkaiah, N, Kiran, Degala Venkata
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.04.2022; Springer Nature B.V
• Subjects: Algorithms; Chemistry and Materials Science; Defects; Materials Science; Metallic Materials; Nondestructive testing; Optical communication; Quality assessment; Review Article; Signal classification; Solid Mechanics; Theoretical and Applied Mechanics; Weld defects; Welding current; Spectroscopy; Arc images; In situ weld defect detection; IR thermography; Multi-sensor defect detection; Welding current and voltage signals; Acoustic emissions
• ISSN: 0043-2288; 1878-6669
• DOI: 10.1007/s40194-021-01229-6

Weld defect detection is a crucial aspect for improving the productivity and quality of the welding process. Several non-destructive methods exist for the identification of defects post weld deposition. However, they only help assess the quality of the component and offer no inputs while the welding process is being performed. Real-time or in situ weld defect detection aids in the detection of defects during the welding process, allowing to take corrective measures or halt the welding to avoid further wastage of time and material. The current paper provides a brief description of various types of weld defects and the commonly used non-destructive testing (NDT) techniques used for identifying weld defects. It then proceeds to provide a detailed review of various methods available for in situ weld defect detection, classifying them based on their input signals. It also classifies the methods based on the type of algorithm used, along with an intuitive explanation of the commonly used algorithms in weld defect detection. The methods covered in this manuscript make use of different input signals that include audio, welding current and voltage, and optical signals also highlighting methods that use a combination of the abovementioned signals for in situ prediction of weld defects. A critical analysis of the efficacy, advantages, and drawbacks of each method is presented. Further, this work highlights a few research gaps identifying avenues for future research in this area.