A New Power Tracker for Distributed and Partially Shaded PV Systems

Renewable energy is gaining ground in distribution networks and is getting widespread throughout the world. This is due to several reasons including a continuous increase in energy demand, reduced supply of conventional fuels, and growing concerns about environmental protection. Electrical energy generated by photovoltaic (PV) power sources is quickly becoming the most promising renewable source mainly due to decreasing manufacturing cost and increased efficiency. Interconnecting a PV source with a load requires a power electronic device made of a DC-DC buck/boost converter and a controller. This device constitutes the so-called maximum power point tracker (MPPT). The proper design of this device is critical for maximizing the output power of the PV module and therefore optimizing the system’s efficiency, especially under varying operating conditions. It is the task of the present paper to tackle the design of distributed MPPT units based on machine learning and artificial intelligence to properly work under these varying conditions. Simulation and experimental results provide a clear picture of the improved performance of the proposed design as compared to state-of-theart techniques found in literature, especially regarding reaction time and power efficiency.