Research on Steelmaking Production Scheduling Defect Monitoring Based on Data Mining Model Identification

Traditional methods are difficult to achieve efficient steelmaking production scheduling and monitoring due to complex data structures and strong hidden defect patterns. This paper proposes a defect recognition model based on multi-stage data mining. Through cleaning and feature engineering, key features such as equipment status and timing logic are extracted from multi-source scheduling data. The DBSCAN clustering and isolation forest algorithms are combined to identify abnormal patterns, and the XGBoost classifier is trained for defect discrimination. Finally, it is embedded in the simulation platform to achieve closed-loop early warning. Experiments have shown that this method effectively reduces the missed detection rate and improves scheduling efficiency. It has good robustness and generalization ability and is suitable for intelligent perception and scheduling optimization in complex industrial scenarios.