Adaptive neural network control for frequency stabilisation in hybrid microgrids

The integration of renewable energy sources and storage devices in modern power systems may result in an islanded microgrid formation at the point of disconnection from the central grid. This offers a great advantage to supply power to critical/ non-critical loads but creates challenges in terms of frequency instability. The paper proposes a neural network-based controller in order to meet the operational stability in the hybrid microgrid having photovoltaic (PV), wind turbine generator (WTG), flywheel energy storage system (FESS), battery energy storage system (BESS), diesel energy generator (DEG) along with an electric vehicle (EV). Initially, the system is designed for frequency stabilisation using PID and Fuzzy PID controllers optimised using moth flame optimisation, whale OPTIMISATION and grey wolf OPTIMISATION algorithm. A neural network-based controller is designed using model reference adaptive control (MRAC) architecture for this paper. The designed adaptive neural network controller provides a more effective means to damp out the frequency oscillations arising during different scenarios in the microgrid. This paper showcases a better and reliable neural network control strategy to mitigate the frequency deviations under a wide range of system operation as compared to the PID and Fuzzy PID controllers.