Nanjing Iron and Steel

Solution

Nanjing Iron and Steel used Huawei Cloud Stack to launch a hybrid large-model factory environment and move AI from pilots into 20 applications across R&D, manufacturing, service, and management. The stack combines large-model and solver capabilities with industrial mechanisms, prediction models, optimization models, domain knowledge, and enterprise data. In production, the platform runs peak-valley power dispatch optimization, coke-blend recommendation, and rolling-force prediction; in R&D it analyzes composition and process data to forecast low-temperature steel performance and adjust parameters online; in quality it uses vision models for metallographic and macrostructure grading; and in office operations it blends knowledge bases with data engines for smart Q&A and traceable analysis. This architecture turns scattered steel data into a reusable AI decision layer that supports modeling, inference, and human workflows across the plant instead of one-off analytics projects.

Data flow

Steelmaking, power, quality, and business data is processed on Huawei Cloud Stack, where large-model and optimization services build forecasting and recommendation models. Those models then feed dispatch, R&D, blending, quality, and Q&A applications back into plant and office workflows for online decision support.