Determinação de parâmetros de módulos fotovoltaicos utilizando o método de Newton Raphson e algoritmos genéticos

Abstract

This research is justified by the expansion of electric power generation through solar plants and by the large-scale adoption of small electric power generation systems, especially photovoltaic (PV) systems using semiconductors connected to the grid. In order to analyze these systems, the reliability of the parameters of these generation systems is very important. The main objective of this work is to study mathematical methods to obtain parameters of photovoltaic modules for static and dynamic studies in medium voltage electric energy systems. Thus, it is necessary to do the mathematical modeling using equations that characterize this type of energy generation system, and when incorporated into the electrical system, its parameters are of great importance. The modeling of photovoltaic cells, the unit component of each module, is done through two models: the one and two diode models. The model with one diode extracts five parameters of the photovoltaic module and the model with two diodes extracts seven parameters. These parameters are: photocurrent, diode reverse saturation current, ideality factor, series and shunt resistances. These parameters are of fundamental importance for the simulation of the power generation system and for the reconstruction of I-V and P-V curves of photovoltaic systems in power system models. In this work, the first method used to extract the parameters was the Newton Raphson (NR), a fast-mathematical method, widely used in electric systems and can be used when one has a very close idea of the solution. Its application is made in the model of one diode and in cases where there is convergence. A Genetic Algorithm (GA) was also used to extract the parameters, which has high optimization capacity and can be used in the solution of nonlinear problems applied to engineering, for a first approximation, when one has no idea of the solution of the problem. In this work the AG is applied in the one and two diode models and has good results in both models, with the two diodes model being indicated in cases of low solar radiation and the model of one diode, indicated for high solar radiation. Finally, AG is used to obtain the initial solution vector for the NR method, joining the two methods to find an optimal point in the solution of the problem. This union allows the use of the Newton Raphson method in all cases in the extraction of parameters of photovoltaic modules and obtain I-V and P-V curves very close to the real curve.