Under complex operating conditions in modern power systems

pre-planned frequency stability control schemes based on strategy tables and implemented through an “offline simulation-online matching” paradigm are subject to a high risk of mismatch. Inappropriate control actions may even trigger secondary disturbances

posing a serious threat to the safe and stable operation of power systems. To address this issue

a response-driven emergency load shedding strategy for frequency stability considering the mismatch impacts of pre-planned schemes is proposed. This strategy is executed after the pre-planned control actions and serves as an effective supplement to them

significantly enhancing system frequency stability. First

a method for calculating the required load shedding amount for frequency stability based on system frequency response (SFR) model identification is established. A sparse frequency measurement-based SFR model identification approach is proposed

upon which an SFR model incorporating stability control is established

and the load shedding amount is iteratively solved according to frequency stability control objectives. Second

a frequency control sensitivity point mining model based on a Transformer network is constructed. By analyzing the mapping relationship between frequency time-series data of key generator bus nodes and frequency control sensitivity points

response-driven online identification of frequency control sensitivity points is achieved. Finally

the total control action is rapidly allocated according to the ranking of sensitive points to construct an emergency frequency stability control scheme. The effectiveness of the proposed method is verified on a practical large-scale AC/DC hybrid power system with tens of thousands of nodes.