An improved regression-based and observation global maximum power point tracker methods

Abstract

Solar photovoltaic energy is a vital renewable resource because it is clean, endless, and pollution-free. Due to the fast growth of the semiconductor and power electronics sectors, photovoltaic (PV) technologies are climbing significant attention in modern electrical power applications. Operating PV energy conversion systems at the maximum power point is essential for getting the maximum power output and raising efficiency. This paper proposes a regression-based Perturb and Observe method to quickly find a global maximum power point, avoiding being stuck in local maxima, likewise analytical and metaheuristic methods. The improved control focuses on the narrowed search areas by linear and non-linear regression analyses using the generated PV model on a flexible Python environment. Furthermore, the method's accuracy is validated in real time under variable temperatures, irradiations, and loads. This method was proven with a hardware implementation. The proposed method is more than 98% accurate and can withstand long-term modelling. The suggested regression-based perturbation and observation method provided a short learning time and easy implementation. Additionally, the dynamic recorded results can be visualized for researchers to utilize efficiently. This paper proposes a regression-based perturbation and observation method to quickly find a global maximum power point tracker, avoiding being stuck in local maxima, likewise analytical and metaheuristic methods. Linear and non-linear regression analyses narrow the search area using generated photovoltaic model on a flexible Python environment. Furthermore, the validated method's accuracy is more than 98%. The suggested method provides a short learning time and easy implementation to withstand long-term modelling. image