Identification of the Photovoltaic Module Dynamic Model via Dynamic Regressor Extension and Mixing

Abstract

This brief deals with the problem of online parameter identification of the parameters of the dynamic model of a photovoltaic (PV) array connected to a power system through a power converter. It has been shown in the literature that when interacting with switching power converters, the dynamic model is able to better account for the PV array operation compared to the classical five-parameter static model of the array. While there are many results of identification of the parameters of the latter model, to the best of our knowledge, no one has provided a solution for the aforementioned more complex dynamic model since it concerns the parameter estimation of a nonlinear, underexcited system with unmeasurable state variables. Achieving such an objective is the main contribution of this brief. We propose a new parameterization of the dynamic model, which, combined with the powerful identification technique of dynamic regressor extension and mixing (DREM), ensures a fast and accurate online estimation of the unknown parameters. Realistic numerical examples via computer simulations are presented to assess the performance of the proposed approach—even being able to track the parameter variations when the system changes operating point.