Application methods and challenges of artificial intelligence in power system stability analysis and control

The “dual-high” characteristics of new power systems have profoundly transformed their operational modes and stability mechanisms, posing severe challenges to the safe and stable operation of power systems. Artificial intelligence (AI) technology, with its robust nonlinear feature fitting capabilities and demonstrated high performance in numerous fields, presents an effective solution to address these challenges. However, its incomplete reliability and black-box nature introduce difficulties in its application to power system stability analysis and control. To clarify the current status, challenges, and prospects of AI in power system stability analysis and control, this paper conducts a comprehensive review and analysis of relevant research. First, the complex safety mechanisms and diversified operational modes arising from the “dual-high” characteristics are analyzed. It is highlighted that traditional methods based on mechanism modeling and offline simulation suffer from real-time limitations and scenario coverage constraints, thereby demonstrating the necessity of AI applications. Second, the current state of AI applications in stability assessment (including angle, voltage, and frequency stability) and stability control (including preventive and emergency control) is systematically reviewed and evaluated. Finally, the core challenges facing practical AI applications are elaborated, including mechanism-integrated learning, model interpretability, security decision-making guarantees, and efficient cognition of complex systems. The research progress is summarized, and corresponding recommendations are proposed.