A Novel Curve-Fitting Algorithm Based on Levenberg-Marquardt Method of Single-Diode Model Photovoltaic Modules

Abstract

Aiming at matrix singularity and decrease accuracy among the processing of curve-fitting algorithm of singlediode
model photovoltaic (PV) modules, this paper proposes a novel curve-fitting algorithm based on Levenberg-
Marquardt Method. According to the sample data with the same voltage sampling interval of monocrystalline
silicon photovoltaic module and polysilicon photovoltaic modules under standard environment, practical results
shows that the proposed curve-fitting algorithm has the advantage of convergence effectively and broad
applicability. Besides, comparing with the previous curve-fitting algorithm based on the Modified Newton method
and Reduced-forms, proposed algorithm has a superior advantage of high accuracy.