Intelligent Analysis Method for Dynamic Response of Embedded DC Power Grids Based on Cluster Analysis

With the integration of high-voltage direct current (HVDC) transmission systems into large-scale alternating current (AC) power grids, typical faults such as commutation failure and DC blocking significantly intensify the coupling between AC and DC systems, resulting in more complex dynamic behaviors and posing serious threats to power system security and stability. To address the difficulty of analyzing complex dynamic responses in AC/DC systems under large disturbances and various operating conditions, this paper proposes an intelligent analysis method for the dynamic response of embedded receiving-end DC grids based on clustering analysis. The method integrates electromechanical-electromagnetic hybrid simulation with machine learning techniques. Principal component analysis (PCA) is employed for high-dimensional feature reduction, and a two-stage clustering model is constructed using density-based spatial clustering of applications with noise (DBSCAN) and K-means algorithms. This enables automatic clustering, identification, and severity assessment of large-scale dynamic simulation data. Furthermore, the method extracts typical dynamic response patterns of AC/DC systems under various fault scenarios and automatically labels and classifies dominant security and stability issues. The effectiveness of the proposed method is verified through case studies based on typical grid operating modes. Simulation results demonstrate that the method can effectively extract representative dynamic behavior patterns under fault conditions, providing strong technical support for the analysis of complex fault mechanisms and the development of security and stability control strategies in hybrid AC/DC systems.