A Fast Method for Generating Security Boundaries in Complex Scenarios

This paper addresses the boundary analysis of the Thermal Hazard Safety Region (THSR) in power grids and proposes a boundary selection model based on AdaBoost.M2. It classifies operating states through DBSCAN clustering, constructs a system boundary set, and trains an ensemble learner using the AdaBoost.M2 algorithm. Verified with the IEEE 39-bus system, the results show that this method performs well in terms of boundary prediction accuracy, especially in scenarios with multiple over-limit lines.