A multi-line linked workshop dynamic scheduling system

Abstract

The application belongs to the technical field of automatic control, and discloses a multi-production line linkage workshop dynamic scheduling system, a data sensing feedback module adopts a dynamic sampling strategy bound with a control instruction execution deviation, normal working condition low-frequency sampling reduces system load, abnormal working condition high-frequency collection of key deviation parameters guarantees timeliness of abnormal data capture; meanwhile, a heterogeneous data normalization model is constructed based on federated learning, under the premise of protecting data privacy, realizing uniformity of ERP, MES and other multi-system data formats, forming a full-process data closed loop with data quality verification and special collection channels, effectively eliminating invalid data, providing real-time, unified and high-quality data sources for dynamic scheduling; a twin modeling module constructs a three-layer architecture multi-production line global digital twin model, considers process and resource coupling relationship between production lines, and real-time corrects model parameters through a deviation feedback mechanism to virtually simulate and predict control deviation.

Images