A Data-Driven Method for Online Gain Scheduling of Distributed Secondary Controller in Time-Delayed Microgrids

Abstract

Secondary control of a microgrid is to restore the frequency/voltage and share power among different units. However, due to time-delay issues in secondary control loops, system instability may happen. To solve this problem, a data-driven approach is proposed for scheduling the control gains to ensure the stability of the microgrid. By utilizing real-time measured current as input, the proposed method can appropriately adjust the control gain setting for the distributed secondary control to achieve a stable operation even under various time-delay scenarios. First, a time-delayed small-signal model is developed for microgrid stability analysis. Based on the damping ratio calculated from the small-signal model, a constrained soft actor-critic (SAC) algorithm is designed to learn an optimal policy of gain scheduling, which can improve the safety and efficiency of learning. Finally, case studies are carried out to validate that the proposed method can provide an optimal gain scheduling policy, which enhances the stability of microgrids during real-time operation.