Optimizing Hybrid Microgrid Performance with Nonlinear Autoregressive Neural Networks in Grid-Connected and Island Mode Operations

Microgrids are becoming more and more significant as the electric energy industry expands into distributed power in recent years. Power quality (PQ) was quite impacted through the incorporation of renewable energy sources into the local distribution system and the nonlinear behaviour of sophisticated power electronic devices. During the micro-grid transition, the main issues are raised i.e., the supply not reached to the loads as per the demand and power quality like voltage instability, less power factor, harmonic distortions, and transients, arise. By using fuel cells, wind turbines, and photovoltaics, a single microgrid is constructed using this suggested topology’s distributed energy resources (DERs). The inverter is controlled using a nonlinear autoregressive exogenous input neural network (NARX-NN) technique to keep the power quality problems within the specified limits and to generate uninterrupted power supply in both grid connected and islanded mode operations.