An adaptive TD3-based control strategy for energy storage systems in integrated wind-solar-storage VSC-MTDC networks

The voltage source converter based multi-terminal DC transmission (VSC-MTDC) system integrating wind-solar-storage is one of the primary solutions for addressing the problem of new energy consumption. Virtual capacitor control is often introduced to enhance the DC-side inertia of VSC-MTDC systems. However, only increasing the virtual capacitance can degrade system damping and trigger instability. To address this, an adaptive control parameter adjustment method based on the double-delay deep deterministic policy gradient (TD3) algorithm and considering small-signal stability constraints is proposed. First, a linearized state-space model of the wind-solar-storage VSC-MTDC system incorporating virtual capacitor control is established. Second, a multivariable coordinated analysis method is employed to determine the feasible small-signal stability region for key control parameters. Furthermore, a parameter adaptive adjustment framework based on TD3 is constructed, utilizing twin Q-networks and delayed update mechanisms to optimize control parameters in real-time within the stability constraints. Simulation results demonstrate that the proposed strategy can effectively compensate for the negative damping effect introduced by the virtual capacitor, significantly improving the dynamic stability and oscillation suppression capability of the system under renewable energy fluctuations while maintaining DC-side high inertia support.