"Catching Defective Products and Improving Processes in Virtual Factories"…Inside the Metaverse Factory [Reading Science]

Experience at KAIST's AI Metaverse Factory Opened on the 1st
Continuous Process Analysis and Defect Detection Possible in Virtual Space Without Stopping the Factory

[Asia Economy Reporter Kim Bong-su] "Welcome to the Manufacturing AI Metaverse Factory."

On the 1st at the KAIST (Korea Advanced Institute of Science and Technology) campus in Daejeon, when I put on the virtual reality (VR) equipment provided by the guide, an empty factory interior appeared. Suddenly, I became a technician wearing a yellow safety helmet and an orange vest. It was the moment I entered the 'Manufacturing AI Metaverse Factory' set up in the virtual world. Just in case, I looked around 360 degrees front, back, left, and right. It felt like I was completely alone, with only the dark factory visible in my field of vision. After a moment, the lights turned on, brightening the factory interior, and a plastic injection molding machine appeared inside the virtual factory. Following the instructions, participants operated the machine and injected fiberglass-reinforced plastic as raw material; the machine whirred and produced one plastic screw.

◆AI + Metaverse = Virtual Factory

The highlight began from this point. The AI Metaverse Factory immediately started judging defective products and analyzing processes by comparing data collected through sensors installed on virtual machines and factories based on accumulated manufacturing data. Assuming about 500 repeated processes were performed, an AI system collected and compared variables such as pressure, speed, position, time, and temperature. Shortly after, the AI Metaverse Factory presented results indicating defects such as bubbles and whitening occurred, noting that the temperature during the injection process was somewhat unstable, based on clearly organized data.

When asked "what this thing is used for," Kim Heung-nam, head of KAIST’s ‘K-Industry4.0’ Promotion Headquarters, defined it as a "virtual smart factory combining manufacturing AI and the metaverse." According to Director Kim, attempts are ongoing to optimize production processes and minimize defects by collecting and analyzing manufacturing data through AI technologies such as machine learning. However, in most cases, applying AI requires stopping the production line. The AI Metaverse Factory, however, can reproduce the process at the same level as an on-site factory and use AI for analysis and simulation. Since errors can be identified and improved during this process, there is no need to halt the factory for process analysis. Especially, whereas technicians have relied solely on their ‘intuition’ until now, the AI Metaverse Factory quantifies and weights various influences and causes affecting production results through scientific analysis, allowing immediate identification and improvement of causes.

◆Work Anytime, Anywhere

The limitations of space also disappear. As long as virtual reality equipment and solutions are available, simultaneous collaboration is possible anywhere in the world. For example, when training on how to operate machines exported to Vietnam or when repairs are needed due to problems, there is no need to go to the site; these can be resolved here at the ‘AI Metaverse Factory.’ This is similar to the principle NASA recently applied by using machine learning AI systems to enable astronauts to repair the International Space Station (ISS) without daily communication with the ground or manuals, significantly reducing time and effort.

While existing VR and AR content only allow first-person experiences, this AI Metaverse Factory faithfully implements an ‘interactive’ platform where users can communicate with others and experience situations identical to reality through virtual reality. It felt like a new dimension beyond one-way systems such as games, meetings, or online lectures and training?AI and metaverse technologies enabling interactive communication in actual production activities have begun to be utilized. At this year’s World Economic Forum (WEF) held at the Swiss Davos Forum, it was agreed that the biggest changes in manufacturing over the next five years will be determined by AI’s machine learning.

The AI Metaverse Factory experienced that day had only one plastic injection machine, but it is said that in the future, it could be applied to factories producing products with complex connected machines and an enormous number of variables. The era has opened where AI in virtual factories detects defective products, improves processes, and saves materials and fuel. Director Kim said, "Large corporations are already analyzing and improving processes using similar technologies," adding, "Our goal is to promote process efficiency using the AI Metaverse Factory for 10,000 small and medium-sized enterprises in the future."