Robots handle molten iron and welding at POSCO; skilled workers now only control them

At POSCO's Pohang Steelworks, the demanding task of removing impurities from molten iron, once a source of extreme physical strain for workers, is now being handled by an AI-driven autonomous operation system. Previously, laborers manually operated a "skimmer" robot via joystick, performing over 8,000 fine adjustments daily, leading to significant musculoskeletal fatigue. The new system, however, reads site data and precisely controls the equipment, allowing for simultaneous adjustments of the skimmer and the ladle, a maneuver previously deemed too risky.

"When there's little impurity left in the molten iron, this system moves the skimmer based on the location and pattern of the remaining impurities, gathers them to one side, and removes them all at once," explained Kwon Oh-hyeong, a manager at the steelworks. He noted that even highly experienced workers with decades of experience have acknowledged the system's capability, stating, "It does as much as I do."

“

When there's little impurity left in the molten iron, this system moves the skimmer based on the location and pattern of the remaining impurities, gathers them to one side, and removes them all at once.

While the system's speed is not yet faster than that of highly skilled human operators, POSCO anticipates its efficiency will increase as it continues to learn from human expertise. The company plans to expand this autonomous operation system to other plants, including the No. 2 steelmaking plant. This adoption of "physical AI", where AI not only analyzes data but also controls physical equipment and robots, is transforming traditional manufacturing sites that relied heavily on human senses and experience.

“

It does as much as I do.