Direct extraction of solar cell model parameters using optimization methods

This article reviews and appraises various fitting parameter extraction methods for solar cells modeled with 5 parameters. The current of illuminated photovoltaic panels and solar cells are customarily described by the single diode model, which contains a diode, a current source, a series resistance and a parallel conductance. This simple model with only 5 parameters provides reasonable accuracy. In this work, we carry out a comparative review of several parameter extraction methods using many data points, since methods based on fewer data points are more prone to measurement error. Among the methods tested in this benchmark are the implicit fitting, the direct optimization of the current (vertical fit) using the Lambert W function, and the Co-content method, which is based on fitting the integral of current with respect to voltage. We also evaluate the direct optimization of the voltage (lateral fit) using the Lambert W function. The results demonstrate that, regarding convergence reliability for a large range of initial values, the best procedure is the Co-content method, which is based on the use of a quadratic equation for fitting. We tested the different methods on the same experimental data of a solar cell as well as PV modules measured by NREL.