Abstract: In response to the challenging issue of effectively filtering constraints in the unit commitment problem constrained by large-scale line transmission, this paper summarizes the operating principles of line constraints in both transient and steady states. It proposes an effective constraint filtering method based on load similarity mining to eliminate ineffective line transmission constraints and simplify problem complexity. Construct distance functions to explore the similarity of historical load data based on the impact levels of different nodes on line flows. Cluster typical power load scenarios according to load similarity and select effective line constraints based on the role of these constraints. Employing a pre-filtering method for system states with long-term unchanged line statuses reduces the computational load in the mining process. Simulations conducted on the IEEE 118 and Case2746wop systems validate the effectiveness of the proposed method and compare it with existing methods, showing that this method efficiently eliminates 99% of ineffective line constraints, reducing solving time by over 80%.

Key words: unit commitment, constraints identification, transmission line congestion, active constraint set, redundant constraints reduction method