Abstract: With increasing renewable energy integration, power systems face security threats such as amplified frequency fluctuations, insufficient inertia, and diminished frequency regulation capability. These challenges increase operational risks and threaten grid stability. Existing frequency constrained optimization methods often struggle to balance flexibility and real-time capability. This paper introduces a multi-dimensional frequency security region construction method based on common-dimensional metrics, defining its boundary by integrating these metrics with weighted coefficients. A model predictive control strategy is applied to the hierarchical optimization framework. By dynamically adjusting generator outputs while accounting for renewable energy uncertainty and multi-dimensional frequency security region constraints, the method achieves real-time feedback and iterative refinement across hierarchical layers, enabling rapid frequency security control. Case studies on IEEE 39 and IEEE 14 systems demonstrate that this model predictive control is based on hierarchical optimization significantly improves frequency stability under high wind power penetration. Results confirm enhanced computational efficiency while ensuring reliable and secure grid operation.

Key words: frequency security, uncertainty, high-penetration renewable energy, security region, unit commitment