Design of a Resistive Brake Controller for Power System Stability Enhancement Using Reinforcement Learning

摘要

Computation of the closed-loop control laws, capable to realize multiple switching operations of a resistive brake (RB) aimed to enhance power system stability, is the primary topic of this brief. The problem is formulated as a multistage decision problem and use of a model-based reinforcement learning (RL) method, known as prioritized sweeping, to compute the control law is considered. To illustrate the performances of the proposed approach results obtained using the model of a synthetic four-machine power system are given. Handling measurement transmission delays is discussed and illustrated.