Visualization of welding phenomena and construction of an online control system.
Keywords: Sensing and visualization of welding phenomena, image processing, AI, deep learning, control, resistance welding simulation, IoT, PLC control.
To visualize the welding phenomenon and achieve high-quality welding, we use visual sensors (CMOS cameras) to capture the welding process, which is then processed through image processing to analyze the phenomenon. We have successfully visualized welding phenomena that are difficult to see due to intense arc light by establishing appropriate imaging techniques. Furthermore, by utilizing AI technology (deep learning), we identify the welding state and control it to achieve good welding results. We are working on building real-time control technology that can be implemented without incurring high costs. In resistance welding, the welding occurs instantaneously, making it difficult to estimate the internal state. Therefore, we conduct numerical simulations using the finite element method (FEM) to realize optimal welding conditions and visualize and analyze complex welding phenomena. We are also actively working to expand the technologies developed for online control of welding phenomena to other fields, enabling us to address challenges not only in welding but also in robot control, IoT, and AI.
- Company：埼玉大学 オープンイノベーションセンター
- Price：Other
