An unmanned crane installed at a POSCO steel mill warehouse loads steel coils onto a vehicle. Photo courtesy of POSCO DX
POSCO DX, the systems integration (SI) affiliate of POSCO Group, has unveiled a blueprint for an intelligent factory business that controls steel industry processes with artificial intelligence (AI) and operates large equipment without human workers.
POSCO DX (022100) held a press conference at the Lotte Hotel in Jung-gu, Seoul, on the 29th, presenting the key status and future plans of its AI transformation (AX) business. The press conference was attended by Choi Tae-hwan, head of the management planning office; Cho Seok-ju, head of the AX Convergence Research Institute; Yoon Seok-jun, head of the Robot Automation Center; and Kim Byung-seok, leader of the AI Workforce Task Force.
The event featured an introduction to the physical AI technology used to build intelligent factories, along with demonstration cases. The two core technologies of the intelligent factory envisioned by POSCO DX are manufacturing-specialized AI and industrial robots. Manufacturing-specialized AI refers to a method of using AI—developed with a focus on the cognition, thinking, and action required to solve specific problems in manufacturing processes—like a field operator. Steel mills have field operators who run various machines and equipment, and the ultimate goal of manufacturing-specialized AI is for the AI to carry out their skilled work know-how in their place. As part of its manufacturing-specialized AI research and development, POSCO DX has secured technologies such as a system for detecting surface defects on carbon steel slabs and autonomous driving technology for torpedo ladle cars (TLC).
“AI monitors operating conditions in real time and shows the flow of logistics within the workplace on an intuitive screen, and we are also developing a conversational AI control system that inputs commands in natural language,” said Cho Seok-ju, head of POSCO DX’s AX Convergence Research Institute.
Cho Seok-ju, head of POSCO DX’s AX Convergence Research Institute, explains POSCO DX’s physical AI business at the Lotte Hotel in Jung-gu, Seoul, on the 29th. Photo courtesy of POSCO DX
Industrial robots involve converting large equipment installed at major processes within the steel mill into unmanned robots. POSCO DX is carrying out work to make three pieces of equipment—raw material unloading grabs, raw material reclaimers, and thick plate cranes—unmanned, and is currently going through test operations. The raw material unloading grab is a tong-shaped device that picks up and unloads raw materials loaded on ships. A single grab weighs as much as 20 tons. POSCO DX is currently conducting research to make 80% of grab operations unmanned. The raw material reclaimer is equipment used to scoop up raw materials piled in outdoor space (yards) and return them into the factory.
The thick plate crane is equipment used to lift thick steel plates 6mm or more in thickness using electromagnets. Thick plates in particular are produced in different shapes according to customer requests. This requires high-difficulty work of lifting and moving irregularly shaped products in units of two or three sheets. POSCO DX plans to formally introduce its industrial robot technology into POSCO steel mills during the second half of this year.
An impurity (dross) removal robot installed at POSCO’s steel mill plating plant. Photo courtesy of POSCO DX
The deployment of robots to build intelligent factories is also underway. POSCO DX regards three matters as key priorities in its robot deployment: automating irregular manual work, automating manual devices, and operating a robot optimization platform. In particular, it is focusing on having robots carry out human tasks that are performed repeatedly in dangerous environments within steel mills. Currently, a robot that removes impurities generated during the plating process is installed at the Gwangyang steel mill. At the Pohang steel mill, robots have been deployed for the repetitive work of rolling steel coils into circles (banding) or unrolling them flat (debanding), successfully achieving automation.
“In addition, POSCO DX is conducting advance research on technology to monitor dangerous situations by deploying quadruped robots within steel mills, and on technology to operate the thin plate bending tester in the material testing room with a humanoid robot,” Yoon said.
At the event, POSCO DX also revealed its intention to expand the physical AI technology it had developed for POSCO Group affiliates into an external business. “The shipbuilding industry has shown interest in the unmanned control technology for thick plate cranes, and POSCO DX has held a technology briefing session,” Cho said. “Once the physical AI technology for internal group business is completed in the second half of this year, we will then enter external businesses.”
