Digital Transformation in Steel Industry: Past, Present & Future

In 1978, China’s steel output was just 31.78 million tons, with per-worker productivity under 15 tons annually. Today, the country produces around 1 billion tons per year, with per-worker output exceeding 600 tons—some advanced mills even reach 2,000 tons per person. Without this leap in labor efficiency, the industry would need over 60 million workers to sustain current output, making wages far exceed sales revenue. This transformation is rooted in advances in equipment, processes, and especially automation—a key enabler that not only replaces manual labor but also catalyzes progress in other technologies.

The Role of PLC and DCS in Steel’s Digital Journey

The programmable logic controller (PLC), a hallmark of Industry 3.0, was adopted early by China’s steel sector. The world’s first PLC appeared in 1969, and when Baosteel began construction in 1978, it integrated PLCs extensively in its first phase. Technologies like continuous casting and converter steelmaking, which replaced ingot casting and open-hearth furnaces, demanded higher automation levels, and PLCs delivered. Distributed control systems (DCS), emerging slightly later, also saw widespread use in Baosteel’s early projects, enabling tighter process control.

These digital tools drove dramatic efficiency gains. In the early reform era, producing one ton of steel consumed 1.2–1.5 tons of coal and 30 tons of water—two to three times and over ten times today’s figures, respectively. Pollution was far worse. With current annual output, without these advances, the industry would burn 1.2–1.5 billion tons of coal, emit over 3 billion tons of CO₂ (a third to a quarter of China’s total), and use 30 billion cubic meters of water—equivalent to the Yellow River’s annual flow into the sea. DCS-enabled automatic control has been pivotal in slashing energy use and emissions.

From Mathematical Models to Industrial Apps

Long before the term “industrial app” became trendy, steel plants used mathematical models—software that bridged management systems and control networks, akin to today’s industrial internet platforms. These models improved steel purity, rolling temperature precision, and dimensional accuracy. By the early 2000s, leading mills recognized such software as core technology, positioning steel ahead in smart manufacturing.

Higher control precision enabled high-end products with ultra-low impurities and tight process windows. Uniform composition can multiply performance, while fine control allows thousands of steel grades and variants. Customization—delivering exact composition, properties, dimensions, and surface quality—adds hundreds of yuan per ton in value. For a 10-million-ton mill, that’s massive revenue. Yet, this variety complicates production planning, inventory, and delivery, especially as order sizes shrink and special requirements grow.

The Data Integration Imperative

When a mill produces 2 million tons annually, manual management can cope. At 6 million tons, it becomes nearly impossible. Customers once faced uncertain delivery times because scheduling involved coordinating countless orders. Excess inventory and “leftover” products (up to 10% in some plants) eroded profits. Computerized management became essential, requiring timely, accurate data from every process. This drove network integration—early efforts at Baosteel used RS485, modems, and even floppy disks to achieve “data without landing” (no human interference in transmission). By the late 1990s, a mainframe computer managed the entire supply chain, marking a new IT height.

Smart Manufacturing: Divergent Paths

As the industry matured, many indicators neared physical limits. The challenge shifted to finding new value. Different regions took different approaches:

• U.S. focus on labor productivity: With GDP per capita at $80,000, U.S. mills must maximize output per worker. Big River Steel achieves nearly 4,000 tons per person by focusing on simpler products and smooth operations, and crucially, by boosting white-collar efficiency through digital tools.
• South Korea’s quality drive: POSCO aims for unmanned smart factories to eliminate human error, achieving unmatched cost and quality competitiveness.

China’s path differs. With per-worker output around 600 tons (and advanced mills at 2,000 tons), the gap is mainly in white-collar and supporting workforce efficiency. The real opportunity lies in leveraging China’s massive, diverse market for customized services and in deep energy/environmental gains. While energy efficiency nears limits, emission reduction has infinite potential—technically, you can always halve pollutants.

Resource Integration and AI-Driven Services

Simpler product mixes boost efficiency. Shandong Yongfeng Steel outperforms many traditional leaders partly due to a focused product portfolio. Digital tools can enable resource integration, allowing each plant to specialize while sharing knowledge and synergies across a group. This multi-party collaboration was identified as a key digitalization focus years ago.

The shift from manufacturing to services is another frontier. High-level steel services traditionally relied on scarce experts, limiting support to large clients. Now, AI and large language models can codify expert knowledge, enabling service for thousands of SMEs. The strategic focus of steel AI should be on market and user service, not just production. Some experts mistakenly apply discrete manufacturing paradigms to steel, ignoring the industry’s unique challenge: uncertainty, not complexity. This misdirection must be avoided.

Key Technologies and Trends

Future Outlook

Digitalization in steel has evolved from basic PLC control to integrated smart systems. The next phase will see deeper AI integration for service customization, resource sharing across enterprises, and relentless environmental improvements. Success depends on asking the right questions—whether it’s boosting white-collar productivity, enabling mass customization, or achieving zero-waste operations. The steel industry’s digital journey is far from over; it’s entering a new chapter where data and AI redefine what’s possible.