Artificial neural network based MPPT and conditional controllers with saturated action for multi-renewable hybrid alternating or direct current microgrids in islanded and grid-connected Modes

Abstract

In the face of energy uncertainty, the microgrid stands as a beacon of resilience, offering a lifeline to communities, businesses, and critical infrastructure. With its ability to seam lessly switch between conventional and renewable energy sources, the microgrid not only ensures uninterrupted power supply but also heralds a future where energy security and sustainability are seamlessly intertwined. Initially, the hybrid AC/DC microgrid incor porates a combination of wind energy, photovoltaics, batteries, supercapacitors, DC-DC power converters, and bidirectional inverters. Subsequent advancements have led to the development of conditional-based super twisting sliding mode controllers (CBST-SMC), which aim to mitigate the wind-up phenomenon occurring in conventional super twist ing sliding mode control when the control signal becomes saturated. These controllers exhibit superior dynamic performance in the presence of external disturbances and ef fectively maintain regulation of the voltage on the DC bus. The focus is also placed on the energy management algorithm implemented within the microgrid. The proposed controller gains are tuned using an improved grey wolf optimization technique, with the objective function being the integral time absolute error. The proof of asymptotic stability of the system is established by applying the Lyapunov stability criterion. To validate the real-time control of the system, an experimental verification using a C2000 Defino microcontroller is conducted through Hardware-in-Loop (HIL) implementation.