Comparative analysis of identification methods of the photovoltaic panel characteristics

For modeling the steady-state characteristic of the photovoltaic panels or cells, particularly the external characteristic current-voltage, I(V), two types of mathematical models can be used: conceptual semi-empirical models, based on electric diagrams, and empirical models, based on pure mathematical reasoning. Each of these models is individualized by a number of parameters that can be obtained by different methods. The present paper concerns with the presentation of three methods to obtain the parameters for three models, one of the first type based on genetic algorithms, and two of the second type. For this purpose, the identification method with genetic algorithm and the methods for the empirical models are summarized. To evaluate and to compare the models and the identification methods a case study is developed. The case study has as starting point experimental obtained characteristics for a photovoltaic cell and a photovoltaic panel. The results gathered for different models by applying genetic algorithm are compared between themselves. These results are compared, also, with those obtained by using the methods for empirical methods. The accuracy of calculated characteristics compared with reference characteristics obtained experimentally and the complexity of calculation methods are used to propose some recommendations for practical use.