Fuzzy Logic Adaptive Crow Search Algorithm for MPPT of a Partially Shaded Photovoltaic System

The arbitrary selection of the Crow Search Algorithm (CSA) parameters, the Awareness Probability (AP) and the Flight Length (fl) results in poor convergence performance and efficiency even if the CSA performs well when solving global optimization problems. In fact, a more search process variety is the outcome of increasing the fl. Furthermore, a higher value of the fl is preferred to guide the optimization process in the direction of global search, whilst a lower fl value directs the algorithm in the direction of local search. In this regard, this study presents a unique Fuzzy Logic adaptive CSA (FL-CSA) for a freestanding Photovoltaic System (PVS) that is based on a Fuzzy Logic (FL) supervisor. Therefore, it is recommended to use the FL supervisor for the online AP and fl, tuning to get superior performance in terms of quick convergence to the GMPP and in terms of high efficiency. Three distinct situations are used to validate the efficacy and speed of the proposed FL-CSA through numerical modeling and experimental testing. The results demonstrate the superiority of the suggested FL-CSA over other traditional approaches, including the Conventional CSA (CCSA), the Conventional Particle Swarm Optimization (CPSO), and the Perturb and Observe (P&O) method. It is true that the maximum power generated by the PVS is extracted by the suggested FL-CSA-based MPPT with average efficiency of 99.93%, whereas the CCSA, the CPSO and P&O record average efficiency of 99.78%, 99.50% and 96.40%, respectively. Additionally, the proposed FL-CSA-based MPPT strategy reduces the convergence time by an average of 42%, 63% and 61%, respectively, in comparison to the CCSA, the CPSO and the P&O MPPT methods.