A Novel Initial Value Prediction of Curve-Fitting Algorithm for Single-Diode Model of Photovoltaic Modules

Abstract

In this paper a new initial value prediction of curve-fitting algorithm is presented for extracting the parameters of the single-diode model. Before extracting the parameters of the single-diode model of Photovoltaic (PV) modules, a reform of the equation of the single-diode model has been presented for improving the accuracy of the Single-diode model. And the biggest advantage of the algorithm is the prediction of initial value where Newton-Raphson method utilized to seek the root of initial value objective function. Also, Gauss-Newton iteration method have been applied to seek the optimal value of the algorithm—the least square nonlinear multiple regression. Test on two classic case studies and operational PV cells and modules, the proposed algorithm exhibits substantial improvement in terms of fitting accuracy, where the root-mean-square error is lower than other methods, and reliabilities for practical applications including different types and materials. With these, the proposed method is envisaged to be valuable for practical applications, in which accuracy of the single-diode model is of prime concern.