Intelligent Low-Bandwidth Frequency Controller for VSGs at Economic Dispatch in Islanded Microgrid

This paper proposes a distributed low-bandwidth intelligent controller (DLBIC) to restore islanded microgrid (IMG) frequency and maintain economic dispatch (ED) among inverter-based distributed generators (IBDGs). The DLBIC utilizes local measurements and receives a single variable from the nearest DG, thereby reducing communication link bandwidth and network infrastructure costs. Notably, the DLBIC is model-free, operating without prior knowledge of the IMG model and parameters. It outperforms other approaches by effectively handling the nonlinearity and complexity of IBDGs without extensive training. Leveraging reinforcement learning principles, the DLBIC can learn autonomously and efficiently manage uncertainties and disturbances within the IMG. The effectiveness of the DLBIC is demonstrated through tests against load connection and disconnection scenarios using the IEEE 38-node benchmark. Results indicate that the proposed intelligent controller successfully restores IMG frequency, mitigates frequency nadir/overshoot, eliminates oscillations, and achieves optimal ED.