Droop-Controlled Bidirectional Inverter-Based Microgrid Using Cascade-Forward Neural Networks

Abstract

The voltage source inverters in microgrids often rely on the droop control method integrated with voltage and inner current control loops in order to provide a reliable electric power supply. This research aims to present a Cascade-Forward Neural Network (CFNN) droop control method that manages inverter-based microgrids under grid-connected/islanded operating modes. The proposed method operates the inverter in a bi-directional technique for a wide range of battery energy storage systems or any other distributed generation systems. The proposed strategy uses the CFNN to learn the inverter’s nonlinear model to achieve accurate demand and reference power tracking under different operating conditions for smart grid applications. Additionally, it reformulates the grid control concept to drive the inverter based on the optimal conditions by considering the power demand, reference power, equipment size, and disturbances. Also, it does not require any tuning procedure. The power tracking and operating performance of the proposed CFNN controller are evaluated through several experimental tests using the power hardware-in-the-loop (PHIL) methodology in different scenarios. All results are matched with the proven conventional strategy to confirm its effectiveness.