Review of explicit models for photovoltaic cell electrical characterization

Abstract

For modeling solar cells, the single diode model presents difficulties in implementation and is expensive computationally because it involves a transcendental and implicit mathematical equation. Some authors have proposed explicit, easy-to-use, and computationally efficient models that approximate its behavior. It is challenging to select the proper model for each specific application because the different proposals were tested for different solar panels, operating conditions, and performance metrics, and, therefore, a direct comparison based on the published information is not possible. In this study, the existing explicit models are reviewed, presenting their equations and discussing their mathematical approximations. Four new models are introduced, and a classification of models is proposed. Furthermore, a comparative analysis of all the models under many photovoltaic technologies and operating conditions is carried out using the same performance metrics and parameter extraction method. This allows developing a framework that makes the selection of models easier for each application. The comparative results show that three models proposed by the research team are more accurate than the implicit approach, with average root mean squared errors as low as 0.41 % (versus 0.54 % error of the implicit model). However, the parameters in these models lack physical sense. Among the explicit models incorporating physical parameters and formulated with elementary functions, the most accurate is based on a first order Padé approximation (0.55 % error). The ranking of models is expected to become a valuable tool for the photovoltaic community in various solar cell modeling tasks.