Extração dos parâmetros do modelo fotovoltaico de um diodo através de um método metaheurístico baseado no algoritmo do morcego

Abstract

This work presents a novel methodology for extracting the one-diode model parameters via experimental data from the characteristic curve in which the photovoltaic generator electrical behavior at steady state is described. Since the optimization problem typically formulated to accomplish this task is continuous, multidimensional, constrained, nonlinear and multimodal in nature, a new metaheuristic optimization technique, fundamentally based on the bat algorithm, has been proposed. This technique, which was named Lévy flight directional bat algorithm (LDBA), gather features brought by a modified version of the bat algorithm, which allow performing a more efficient intensification in the searching process for the optimal solution, enabling to handle it conveniently with the discrepant magnitude orders normally observed between the values assumed by the one-diode model parameters. In addition to these features, a randomization strategy inspired by the Lévy flight and a dynamic procedure for correcting any solution found which violates the established parameters bounds, are incorporated in an innovative way to this technique in order to improve its ability to exploit the solution space problem. The validation of proposed metaheuristic technique is performed from the application of the one-diode model in the representation of the electrical behavior of three photovoltaic generation devices commonly adopted in the literature. Firstly, an analysis is performed specifically to ascertain the real relevance of exploiting the features incorporated into the LDBA. Secondly, a comparative analysis of the solutions quality obtained from this developed technique, as well as its computational performance with existing methods in the literature is presented. Although the parameters extraction of the one-diode model via experimental data has been investigated for decades, it is still an open problem in the photovoltaic industry, being one of the most attractive and fundamental research themes for investigation. Thus, the main contribution of this work is the proposition of a new tool which allows the effective and reliable accomplishment of this task.