Emergency Control for Sending-end Power System with Renewable Energy under HVDC Blocking Based on Deep Reinforcement Learning

To protect the sending-end power system from over-frequency insecurity, a deep reinforcement learning (DRL)-based approach is proposed to compute the remedial action scheme against HVDC blocking in real-time. Firstly, the optimization model for emergency frequency control is formulated. Considering that this optimization model is a nonlinear combinatorial problem with dynamic constraints, deep reinforcement learning-based solution approach is proposed. Case study on the modified IEEE 39-bus system is presented to demonstrate the effectiveness of the proposed approach.