LEO Satellite Internet Assisted Data driven Predictive Control for Remote Island ed Microgrid Stability Enhancement

Grid-forming converters not only provide voltage and frequency support for remote islanded microgrids but also induce increasing instability risks during disturbances caused by typhoons, rainstorms, etc. With the assistance of LEO satellite internet, a data-driven predictive control (DPC) strategy is proposed to enhance the stability and resilience of remote islanded microgrids against widespread and frequent disturbances. Owing to the high-speed and wide-bandwidth communication services provided by the LEO satellite internet, the proposed DPC can establish timely and reliable data interaction among participating converters with acceptable latency, and then a two-layer gate recurrent unit (GRU) neural network is applied to predict the converters' dynamics precisely despite the parameter mismatch drawbacks of mathematic models. Furthermore, the proposed DPC achieves the derived consensus dynamically among converters such that the coherence and consistency are guaranteed effectively. Extensive case studies on practical remote islanded microgrids have been conducted to validate the effectiveness of the proposed DPC in coping with various types of disturbances and oscillations, effectively ensuring the system stability with considering the delays induced by the LEO satellite internet