Construction of Static Security Region based on Data-driven Surrogate Model of Power Flow Equations

Conducting real-time online security analysis for power grids and determining the system’s security margin are pivotal research areas in the field of power systems. The thermal security region plays a crucial role in evaluating how random fluctuations in power growth direction affect the thermal stability of power systems. By employing the DC power flow model, the high-dimensional boundary surface of the steady-state security region can be approximated as a hyperplane. However, as the scale of the system expands, the inaccuracies in the analytical model increasingly distort the security region boundary. To tackle this challenge, this paper presents a hybrid approach that integrates physical model-driven and data-driven methods to modify the steady-state security region boundary. This approach not only retains essential information from the physical model but also harnesses the power of data analysis to uncover hidden linear error patterns. Extensive testing on various IEEE standard test systems demonstrates the model’s exceptional accuracy.