Domestic researchers have developed a module that can be combined with existing factory production equipment to implement artificial intelligence (AI). When the module is attached, it can autonomously recognize defects and find optimal conditions.
A joint research team led by Yoo Se-hoon, senior researcher at the Korea Institute of Industrial Technology, and Lee Ho-jin, senior researcher of the mobility components group at the institute, announced on the 30th that they have developed AI-based metal 3D printing technology capable of autonomously recognizing and improving product defects.
This technology, named "Metal 3D printing defect detection and active control technology," can be used by combining an "add-on" module with aging production equipment that does not have AI applied. The add-on is a module that gathers sensors, defect detection, and equipment control technologies to intelligentize production equipment. By using the add-on module, it is possible to transform existing factory production equipment into an AI-based smart factory without replacing it.
The research team applied the add-on module to direct energy deposition (DED) 3D printing. DED refers to a method in which a high-energy laser is shot at metal powder or wire to melt it and stack layers. In DED 3D printing, defects can easily occur if errors arise in factors such as laser output, stacking speed, and powder supply.
The add-on developed by the research team detects anomalies during the 3D printing process and alerts operators through a display. It also actively controls the process elements to derive optimal conditions and autonomously resolves issues. While operators rely on experience, taking a long time to find the best process conditions, the add-on can automatically recognize, control, and improve product defects, making it usable for manufacturing corporations lacking advanced equipment and specialized personnel.
The research team has also successfully transferred the technology developed this time to domestic corporations MR Tech and Dico for commercialization. Senior researcher Lee Ho-jin noted, "The technology detects various stacking defects using deep learning technology and actively controls the conditions of 3D printing equipment in real time," adding, "It is expected to have a significant impact as it can also be applied to implement digital twin virtual models of production process data."
Moon Chang-kyu, head of MR Tech, said, "It can also be applied to robot-based production processes, and we are promoting the commercialization of AI-based robot 3D printing technology." Dico CEO Hwang Jun-cheol noted, "It has a high competitive advantage in terms of being able to obtain and manage temperature data of production processes based on video systems," and said, "We plan to apply it in aerospace, medical, and automotive fields."