The centralized control system of thermal power plants puts forward higher requirements for the dynamic responsiveness of control strategies. The article studies the functional structure and collaborative control mechanism of DCS and SIS systems

analyzes the state identification method and strategy optimization path based on deep learning

and finally discusses the performance evaluation results of control strategies under typical working conditions. The research results can provide data-driven control support for the automated operation of centralized control systems under complex working conditions

and improve security and response efficiency.