Research on Data-Driven Intelligent Adjustment Technology for Intraday Tie-Line Schedule

As the global energy transition to clean energy accelerates, enhancing the consumption of large-scale gridconnected clean energy has become a key research focus in power system and operation. The traditional tie-line scheduling method driven by physical models has two major limitations: lack of self-learning ability and low efficiency of physical solution, which makes it difficult to deal with the complexity of multi-level planning models in national provinces. This paper proposes an intelligent adjustment scheme driven by data, and constructs an intraday tie-line scheduling model for the national dispatching and provincial dispatching based on the long short-term memory deep neural network. By training historical data to establish mapping relationships, the model acquires self-evolution and learning capabilities, ultimately achieving intelligent adjustment of tie-line schedules.