NANNING, May 4 (Xinhua) — What transformation can emerge when traditional heavy industry meets artificial intelligence (AI)? In south China’s Guangxi Zhuang Autonomous Region, an established steel plant has fully embedded the Xuantie AI model into its entire production process, leading to revolutionary changes.
Departing from the typical image of old steel plants shrouded in dust and noise, Guangxi Iron and Steel, a subsidiary of Liuzhou Steel Group, is clean, tidy and well-organized.
Operating smart equipment precisely and streaming data in real time have become the norm, and AI has been taking part in production control and scheduling, including scenarios such as intelligent ore blending and raw material optimization, smart transportation of molten iron, and predictive equipment maintenance.
At the ironmaking station, molten iron is precisely transported by autonomous trains to designated locations without any human intervention throughout the process.
The plant currently operates three blast furnaces, 10 locomotives and 51 molten iron tank carriages, with each locomotive transporting approximately 3,600 tonnes of molten iron per day, according to Mo Keyi, a train operator.
Since this intelligent molten iron transport system on the two-kilometer-long route was put into operation, the temperature drop of the molten iron has been reduced by 35 degrees Celsius, saving approximately 60,000 tonnes of standard coal and cutting carbon dioxide emissions by 262,000 tonnes annually.
On a thick steel plate production line, red-hot steel plates are repeatedly rolled and cooled to gradually become thinner. Even dozens of meters away, the intense heat can still be felt.
In the past, plate nesting relied heavily on manual input from engineers, said Chen Peizhi, head of modeling technique at the plant, noting that finding the optimal solution among a vast number of combinations was difficult and time-consuming, leading to low raw material utilization.
“Now, the intelligent plate nesting algorithm powered by the Xuantie model transforms the manual experience into a model that can compute the plate nesting solutions for over 1,000 orders within minutes,” Chen said.
On the smart cold rolling production line, robotic arms featuring 5G and AI, intelligent heavy-load transport robots and unmanned vehicles work together, enabling fully automatic, high-precision and remote control of heavy materials.
There is another important step in the steelmaking process known as “flame monitoring by AI,” said Qin Jun, a manager at the manufacturing management department of Guangxi Iron and Steel.
High-definition cameras capture the color, shape and brightness of the flames inside the furnace. After rapid processing by AI algorithms, the system can accurately determine the conditions within the furnace, providing real-time guidance for the steelmaking, said Qin.
The Xuantie model, launched on March 31 and jointly developed by Liuzhou Steel Group, Huawei, China Mobile Guangxi and other partners, is the first industry-specific large model for the steel sector in Guangxi. It has been applied in key areas of steel manufacturing, including raw material purchase, ironmaking and steelmaking, steel rolling, quality management, equipment maintenance and supply chain coordination.
By applying AI technology, Liuzhou Steel Group has achieved an 8.5 percent increase in production efficiency and reduced the cost of producing crude steel by 5 yuan per tonne, said Shen Min, deputy general manager of the group.
Meanwhile, the steelmaker is optimizing its product mix through burgeoning intelligence initiatives. The share of steel products for manufacturing purposes has jumped from 30 percent at the beginning of the 14th Five-Year Plan period (2021-2025) to over 70 percent at present.
The intelligent transformation aligns with the country’s government work report this year, which for the first time stressed the need to create “new forms of smart economy” and to continue deepening and expanding the “AI plus” initiative.
It also calls for the faster adoption of new-generation intelligent terminals and AI agents, large-scale commercial applications of AI in key sectors and fields, and the cultivation of new forms and models of AI-native business.
“We plan to achieve over 80 percent intelligent coverage across the whole production process within three years,” said Li Bin, chairman of the group.
The application of AI technology has not come at the expense of labor interests at Liuzhou Steel Group, which has approximately 27,000 employees. In response to the intelligent transformation, it has taken systematic measures to help workers adapt to new technologies and new roles.
The steelmaker has launched low-code training programs and AI application competitions, ensuring that AI is no longer the exclusive domain of a few tech-savvy employees but has become a productivity tool for the entire workforce, said Lu Zhaogang, general manager of the digital and intelligent development center of Liuzhou Steel Group.
Its employees have spontaneously created over 2,000 AI assistants and more than 16,000 customized applications, covering a wide range of scenarios such as efficiency improvement, cost control, inspections and predictive maintenance.
All these efforts have facilitated the workers’ gradual shift from traditional manual operators to designers and maintainers of the technological system, said Lu. ■