Cold rolling production primarily comprises processes such as pickling, rolling, annealing, and inspection. This production environment involves numerous process constraints; rolling forces can exceed 3,000 tons, and the steel strip travels at speeds of up to 1,200 meters per minute-meaning that even minor fluctuations in any parameter could result in the entire coil being scrapped. In the galvanizing workshop, AI models automatically adjust air knife height and pressure, achieving a "3-gram hit rate" (where the deviation of the zinc coating weight per square meter from the target value falls within ±3 grams) exceeding 99%.
These advancements have resolved critical issues-such as welding instability, a high propensity for strip breakage, and rolling mill vibration-thereby enabling the efficient and high-quality rolling of ultra-high-strength steel. Furthermore, we have independently developed high-precision control technologies for the shape and dimensions of ultra-high-strength steel. We introduced a multivariate coordinated optimal control model-based on thickness and mass flow rate-for full-stand performance prediction and compensation, alongside a predictive control strategy. This system has achieved a 98.77% rate of thickness control accuracy within a tolerance of ±1%, while plate shape accuracy has reached 5 IU.