AI Welding and Robotic Dogs Enhance Manufacturing in South Korea

Collaborative robot for block assembly at HD Hyundai Heavy Industries. [Photo=Joint Coverage Group]

"When manufacturing 'lugs,' essential for moving ship blocks, six workers could produce about 100 units a day manually. Recognizing that this is the only product that can be mass-produced in shipyards, we established an autonomous manufacturing system for lugs using robots, significantly enhancing productivity. The need for artificial intelligence (AI) is growing to maintain competitiveness and increase productivity," said Yoon Dae-kyu, an executive at HD Hyundai Heavy Industries.
As domestic manufacturing faces challenges from intensified price competition from China, an aging workforce, and global supply chain restructuring, AI is emerging as a solution. The transition to manufacturing AI (M.AX) is becoming a necessity rather than an option.
During a meeting on June 12 at HD Hyundai Heavy Industries' medium-sized ship division in Ulsan, Yoon noted that industrial robots were continuously welding and producing or recycling lugs, which are essential components in shipbuilding.
Lugs connect blocks to lifting equipment when cranes lift or move ship blocks. They are produced in various specifications but are used in large quantities throughout the shipbuilding process, necessitating a production system that can supply a variety of lugs in a timely manner. Additionally, since lugs can be reused two to three times, a recycling and management system is also essential.
HD Hyundai Heavy Industries has established a lug autonomous manufacturing system based on eight industrial robots and two autonomous mobile robots (AMRs). This system has transitioned from a manual welding-centric production method to an unmanned production system. According to Yoon, this shift has improved the continuity and stability of the production flow.
Production efficiency has also increased. Since implementing the lug autonomous manufacturing system, production has improved by 87.5%. The automation equipment performs repetitive tasks reliably, enhancing production efficiency and allowing for flexible supply of various lugs. Variations in worker skill levels have decreased, reducing the physical strain on workers and minimizing the risk of industrial accidents.
The use of collaborative robots is also increasing on-site. In the second shipbuilding plant, welding collaborative robots are employed during the assembly of flat blocks. Previously, repetitive welding tasks in confined spaces posed significant risks and discomfort for workers, as well as the potential for musculoskeletal disorders.
These issues are being addressed by collaborative robots that incorporate the expertise of skilled workers. Each robot performs the work of two skilled workers with 5 to 10 years of experience, resulting in a productivity increase of about 70%, according to HD Hyundai Heavy Industries.
Looking ahead, the challenge lies in developing unstructured AI technology. Yoon stated, "Currently, we can handle structured components to some extent, but unstructured components vary by design and product. We are developing humanoids that can be utilized in the dock for ship construction, not only for parts inside the ship but also externally using AI."

Boston Dynamics' Spot robot inspecting a tuyere at POSCO's Pohang Steelworks. [Photo=POSCO]

M.AX is being utilized not only in shipbuilding but also in the steel industry. POSCO is introducing autonomous robot technology for predictive maintenance and high-risk tasks in steelmaking processes. Predictive maintenance involves collecting data and using AI to monitor the condition of machinery in real time to predict failure points.
A notable example is the use of Boston Dynamics' Spot robot to inspect tuyeres in the second blast furnace at Pohang Steelworks. Inspecting the external temperature and gas leaks of the 30 tuyeres that blow air into the furnace is essential. However, with a limited number of workers managing the entire furnace, regular inspections have been challenging. The risk of burns or gas exposure is significant due to the temperatures exceeding 1,100 degrees.
To address this, the company is deploying robotic dogs for tuyere inspections based on accumulated data. The robots are equipped with anomaly detection capabilities through data analysis, enabling real-time monitoring. This data is continuously accumulated to implement monitoring functions based on a digital twin framework.
AI is also expected to be utilized for inspecting rollers on belt conveyors that transport steel and for manual steelwork. By detecting anomalies based on audio data from the belt conveyor, robots can be deployed for replacements. The plan is to minimize the deployment of workers in high-risk areas by having humanoid robots perform tasks previously done by humans near molten metal.
The technologies developed through this initiative are expected to be applicable in similar industries in the future. Choi Yong-jun, a researcher at POSCO, stated, "After enhancing the diagnostic performance of key equipment, we plan to expand robot demonstrations and create an integrated platform for predictive maintenance packages to facilitate technology transfer."

* This article has been translated by AI.