Global Maximum Power Point Tracking of Photovoltaic Systems Using Artificial Intelligence

Recently, artificial intelligence (AI) has become a promising solution to the optimization of the energy harvesting and performance of photovoltaic (PV) systems. Traditional maximum power point tracking (MPPT) algorithms have several drawbacks on tracking the global maximum power point (GMPP) under partial shading conditions (PSCs). To track the GMPP, AI enabled methods stand out over other traditional solutions in terms of faster tracking dynamics, lesser oscillation, higher efficiency. However, such AI-based MPPT methods differ significantly in various applications, and thus, a full picture of AI-based MPPT methods is of interest to further optimize the PV energy harvesting. In this paper, various AI-based global maximum power point tracking (GMPPT) techniques are then implemented and critically compared by highlighting the advantages and disadvantages of each technique under dynamic weather conditions. The comparison demonstrates that the hybrid AI techniques are more reliable, which offer higher efficiency and better dynamics to handle PSCs. According to the benchmarking, a modified particle swarm optimization (PSO) GMPPT algorithm is proposed, and the experimental results validate its ability to achieve GMPPT with faster dynamics and higher efficiency. This paper is intended to motivate engineers and researchers by offering valuable insights for the selection and implementation of GMPPT techniques and to explore the AI techniques to enhance the efficiency and reliability of PV systems by providing fresh perspectives on optimal AI-based GMPPT techniques.