POSCO steel mills enhance production process with AI

2024. 8. 26. 11:27
글자크기 설정 파란원을 좌우로 움직이시면 글자크기가 변경 됩니다.

이 글자크기로 변경됩니다.

(예시) 가장 빠른 뉴스가 있고 다양한 정보, 쌍방향 소통이 숨쉬는 다음뉴스를 만나보세요. 다음뉴스는 국내외 주요이슈와 실시간 속보, 문화생활 및 다양한 분야의 뉴스를 입체적으로 전달하고 있습니다.

The AI CCTV system was implemented at the 4th Continuous Casting Plant at Pohang Steelworks. [Courtesy of POSCO DX]
South Korea’s POSCO Group is expanding the use of artificial intelligence (AI) technology throughout its steel mills to streamline steel product production processes and enhance worker safety.

During a recent visit to the Pohang Steelworks No. 4 Casting Plant during the previous week, AI-equipped smart closed-circuit TVs were observing slabs moving on rollers.

A slab is a semi-finished steel product made by solidifying hot molten steel from a blast furnace. They are cut by two torches located at both ends and then move along rollers once they solidify. During this movement, if there are foreign substances on the cut surface or if the slab is not aligned properly, the slab could deviate and not move in a straight line, which happens about five to six times a month.

Slabs are about 8 meters long and weigh between 30 and 35 tons. If a slab moving diagonally is not corrected, it could lead to collisions that destroy equipment and cause the entire factory to. The slabs’ temperatures can hit the 1,000 degrees Celsius range, which makes repositioning a colliding slab challenging and detecting distortions in advance crucial.

POSCO DX introduced AI functionality to the smart CCTV system at the No. 4 Casting Plant starting in 2022. The CCTV monitors the slabs’ movement in real time and pauses the equipment if there is a risk of collision due to misalignment. No accidents due to slab misalignment have occurred since the AI solution was implemented to date.

The company has also applied AI to automate the inspection process of wire rod products shipped to customers, with 12 smart CCTVs in the inspection area working to find labels attached to the products. When a truck loaded with 14 coils, each weighing 2 tons, arrives at the inspection area, the system first identifies the vehicle’s license plate. The smart CCTVs then locate the labels on the coils, use the zoom functions to discern the information, and compare it with the invoice.

POSCO DX also uses vision AI technology on the tracks of 1,200-ton locomotives that transport molten steel. This technology helps by sending alerts to train operators if people or vehicles enter railway crossings, allowing them to reduce their speed.

Copyright © 매일경제 & mk.co.kr. 무단 전재, 재배포 및 AI학습 이용 금지

이 기사에 대해 어떻게 생각하시나요?