BaoWu Steel Group

In the hot rolling production line, before a steel billet with a thickness of 260mm is rolled into a steel plate with a thickness of 1.2mm, it needs to go through 20 processes, involving over 300 parameters. Each time the type and size of the steel plate need to be adjusted, engineers have to reconfigure the settings, which takes about 5 days each time.

The Pangu Large model has been deployed in the 1880 hot rolling production line of Baowu Steel Group, improving prediction accuracy by over 5% and increasing the steel plate yield rate by 0.5%. It is estimated that an additional 20,000 tons of steel plates can be produced annually, generating an annual revenue increase of over 90 million RMB.

Through precise blast furnace temperature prediction, steelmaking enterprises can better control the smelting process, optimize in-furnace reaction conditions, and reduce energy consumption and raw material waste. Additionally, it can improve the consistency and reliability of the products, thereby enhancing the competitiveness of the enterprise and laying a foundation for sustainable development.

The Pangu Large model has an accuracy rate of 88-90%, significantly improving prediction precision compared to the 83-85% accuracy of machine learning models, achieving a substantial enhancement.