Due to the uncertainty in energy transition and the bidirectional flow of power between generation and load

voltage control in power systems has become more complex and challenging. In response to this issue

a data-driven voltage stability margin optimization control method is proposed

which defines the voltage stability margin by deriving the physical quantities of power flow in power grid and constructs a voltage stability margin control model. The model derives the voltage stability margin from power flow data and builds its control model. By establishing a Markov decision process and using the twin delayed deep deterministic policy gradient (TD3) algorithm

the real-time control actions for voltage stability margin of the grid are obtained. Simulation results show that this model can respond quickly to systems and has significant advantages in networks with load fluctuations

which can enhance the voltage stability margin and ensure high-quality power supply.