Data-driven Modeling of Commercial Photovoltaic Inverter Dynamics Using Power Hardware-in-the-Loop
Nischal Guruwacharya, Harish B. Bhandari, Sunil Subedi, Jesus D. Vasquez‐Plaza, Matthew Lee Stoel, Ujjwol Tamrakar, Felipe Wilches‐Bernal, Fabio Andrade, Timothy M. Hansen, Reinaldo Tonkoski
Abstract
Grid technologies connected via power electronic converter (PEC) interfaces increasingly include the grid support functions for voltage and frequency support defined by the IEEE 1547-2018 standard. The shift towards converter-based generation necessitates accurate PEC models for assessing system dynamics that were previously ignored in conventional power systems. In this paper, a method for assessing photovoltaic (PV) inverter dynamics using a data-driven technique with power hardware-in-the-loop is presented. The data-driven modeling technique uses various probing signals to estimate commercial off-the-shelf (COTS) inverter dynamics. The MATLAB system identification toolbox is used to develop a dynamic COTS inverter model from the perturbed grid voltage (i.e., probing signal) and measured current injected to the grid by the inverter. The goodness-of-fit of COTS inverter dynamics in Volt-VAr support mode under each probing signal is compared. The results show that the logarithmic square-chirp probing signal adequately excites the COTS inverter in Volt-VAr mode to fit a data-driven dynamic model.