Advancements in Intelligent Robotic Arc Welding: A Comprehensive Review of AI and Sensor Integration

This review focuses on the development of arc welding into intelligent robotic arc welding (IRAW) systems. It discusses (1) the part and integration of different sensor technologies, (2) Artificial Intelligence (AI) and machine learning (ML) in controlling the process and detection of flaws, (3) the current limitations of IRAW systems, and (4) the prospects of IRAW system development. The IRAW system combines welding environment sensors, process sensors, and fault detection for the actual control of the real-time mode of the process. The penetration of the welds, width and porosity/inclusion flaws, and temperature change were measured using mounted cameras, fast-motion cameras, and infrared cameras. The integration and acquisition of data and the use of big data analysis have also been employed to identify the most favorable welding environment or decrease the defect rate. Sensors and algorithms that are used in robotic welding systems are becoming more prevalent. They assist in improving and accelerating welds because they regulate the welds in real-time and detect defects automatically. This has been beneficial in most industries. New welding processes and closed-loop control systems have been developed to modify the material, shape of the joints, and environment to create weldments of uniform quality. Machine learning models like CNNs, RNNs, LSTM/GRU, CNN-LSTM, and ANFIS, along with tools like Grad-CAM, can identify problems. This helps improve control systems, reducing defects and boosting production. Sensing technologies, adaptive control, and artificial intelligence are the future of the welding industry because they will present high-quality, more productive, and sustainable features to the welding industry.