POSCO uses AI vision to cut waste in steel production process

POSCO Pohang plant’s third steel factory [Photo provided by POSCO DX]

POSCO, South Korea’s biggest steelmaker, is leveraging artificial intelligence (AI) vision to reduce waste in its steel manufacturing process as manual processes to remove impurities such as slag in molten iron from a blast furnace, results in a loss of about 10 percent of the molten iron.

The company aims to expand the installation of AI cameras in its steelmaking division, which is expected to save billions of won in annual costs. Workers at Pohang plant’s third steel factory, which introduced AI cameras in March, expressed their anticipation of the new technology to Maeil Business Newspaper during a recent visit.

POSCO posted a record revenue of 84 trillion won ($63.54 billion) last year, with double-digit growth. However, due to the steel industry recession this year, its growth rate contracted in the first quarter from a year ago. Therefore, the company requires productivity improvements even more urgently.

In this situation, the AI cameras that help save molten iron, which is an essential element in steel production, have been proven to be effective, and POSCO workers here are optimistic.

The third steel factory has a large scoop-shaped skimmer that moves busily to remove impurities from the molten iron. Within the work area where temperatures can reach 1,300°C, four AI cameras are installed, with two on the 5-meter-high ceiling and the two others diagonally on the side.

The AI cameras are surrounded by camera covers and coolers to maintain an operating temperature of 20-30°C. The cameras capture real-time images of the container holding the molten iron and impurities. The impurity images displayed in black on the screen are transmitted to the control room. The skimmer then automatically removes the black material in the image through the AI vision system.

POSCO plans to install these AI cameras in six factories. The AI cameras are based on vision AI technology provided by POSCO DX Co., an information technology service company under POSCO Group. Steel production via blast furnace typically involves three stages: iron making, steelmaking and rolling. In the steelmaking process of removing impurities, about 10 percent of the molten iron is estimated to be lost.

At the Pohang steel plant’s third steel factory, 300 tons of molten iron are processed 45 times a day, resulting in the loss of about 30 tons of molten iron during the removal of impurities. This loss requires supplementing with additional scrap or ingredients, which can cost over 300,000 won per ton.

However, the AI cameras installed at this plant reduce these losses. For every 300 tons of impurities removed, the plant can save about 1 ton of molten iron. This is worth billions of won per year.