Improved Adaptive Control Method of Virtual Synchronous Generator for Enhancing Transient Rotor Angle Stability of New Power Systems

Abstract

With the increasing integration of large-scale renewable energy sources (RESs), modern power grids are evolving into low-inertia and weak-damping networks, creating challenges for transient power-angle stability. Conventional grid-following (GFL) control strategies present a lagging synchronization response under weak grid conditions due to their reliance on phase-locked loops (PLLs). Although grid-forming (GFM) control via virtual synchronous generators (VSGs) provides standalone voltage source properties, conventional fixed-parameter VSGs exhibit an inherent design trade-off between first-swing angular suppression and post-fault oscillation damping during severe short-circuit disturbances. To resolve these vulnerabilities, this paper proposes an adaptive VSG control strategy designed to improve the transient power-angle response of the investigated system. By establishing a parametric judgment framework based on real-time frequency deviations and acceleration rates, the virtual inertia Jt and damping coefficient Dt are adaptively adjusted within predefined limits. Furthermore, a current-limiting mechanism is incorporated into the parameter adaptation loops to prevent converter overcurrent tripping. To evaluate the proposed method, an aggregated grid-connected benchmark system comprising a 5 MVA PMSG-based wind farm, a conventional synchronous generator, and an external grid is implemented in MATLAB/Simulink. The simulation results under continuous ambient operational fluctuations and severe three-phase short-circuit faults demonstrate that the proposed strategy reduces the first-swing power-angle peak by approximately 31% compared with the conventional fixed-parameter VSG, and the response settles within approximately 0.4 s in the investigated fault case. These results indicate the simulation-based feasibility of the proposed bounded adaptive strategy. Further real-time, hardware-in-the-loop, and experimental validation is required before practical implementation.