Dynamic Polynomial Models With ARMA Coefficients Used for Modeling the DC Series Arc Fault in Photovoltaic Systems

DC series arc fault (DCSAF) in photovoltaic (PV) systems is a controversial issue owing to the challenges related to its modeling and protection. By utilizing the actual recorded data, for the first time, polynomial models are used to model the DCSAFs in PV panels. In the models, the nonlinear, dynamic, and time-variant behavior of the DCSAFs are taken into account. In addition to the nonlinear terms, the polynomial models also include derivative terms to show the dynamic nature of the arc. To determine the suitable model orders, various models are studied and the model with the minimum value of error index is selected as the best model. Unlike the well-known traditional models, the coefficients of the polynomial models are time-variant, which is assessed using the autoregressive moving average process. By the actual recorded data, the performance of the polynomial models is evaluated. The criteria for comparing and selecting the polynomial models rather than other traditional models are the error indices. Analyses show that the polynomial models are more efficient to model DCSAF in comparison with the famous traditional models.