Dynamic Global Maximum Power Point Tracking for Partially Shaded PV Arrays in Grid-Connected PV Systems

Abstract

Global maximum power point tracking (GMPPT) algorithms can extract the maximum available power from photovoltaic (PV) arrays even under partial shading conditions (PSCs). The existing GMPPT algorithms originate from computationally-intensive optimization (heuristic) or artificial intelligence concepts, which operate in discrete time steps and impose intense variations to the demanded PV array voltage/current. These result in undesirable disturbances, which increase the overall time required for the GMPPT process to complete and affect the quality of power injected to the grid. In this article, a new GMPPT method with low computational complexity is presented, which exploits the dynamic response of the PV system. The proposed GMPPT technique can track the GMPP in significantly less time when applied to PV inverters with high PV-side capacitances, guarantee convergence to the GMPP even under complex PSCs, while also avoiding the aforementioned disturbances. The performance of the proposed GMPPT method is evaluated using an experimental setup incorporating a 2-kW single-phase grid-tied transformerless PV inverter and a rooftop PV array. The experimental results show that it can identify the GMPP in approximately 1 s under various operating conditions, which is more than 95% faster than the power-voltage curve scanning and particle swarm optimization GMPPT algorithms.