DEFECT DETECTION IN WIRE ARC ADDITIVE MANUFACTURING THROUGH REAL-TIME CURRENT MONITORING

Abstract

Wire arc additive manufacturing (WAAM) is a promising technique that can be used to build complex, large-scale metal components. It offers several advantages over traditional production techniques, such as less material waste, quicker lead times, and increased design freedom. Like any manufacturing process, WAAM has its share of challenges. Errors can happen during the additive manufacturing process, jeopardizing the integrity and quality of the finished product. In WAAM, porosity is one flaw that frequently occurs. It may significantly reduce the mechanical and fatigue resistance of the resulting component. To address this defect, it is essential to modify the process parameters, such as the shielding gas flow rate, wire feed rate, and heat input, in order to produce proper fusion and limit the occurrence of porosity. Understanding these issues and implementing effective mitigation techniques are critical to the successful implementation of WAAM. Machine learning has emerged as a practical tool for defect detection in the WAAM process as a result of the limitations of previous approaches. Machine learning models can identify flaws with precision and tact by using high-quality training data and preprocessing techniques. However, problems like the lack of labeled training data and the requirement for real-time model implementation still need to be addressed before machine learning may be easily combined with other defect detection methods in the future.