A Passive Islanding Detection Method of Grid-Connected Photovoltaic System based on Data-Driven Subspace Identification

The rapid advancement of photovoltaic grid-connected power generation technology has presented notable challenges, including the occurrence of the islanding effect. The structure of distributed photovoltaic power generation involves the interconnection of individual inverters, which poses challenges for traditional islanding detection methods designed for single inverters. As a result, the detection of islanding events becomes more complex and less effective in this context. Therefore, this paper presents a novel approach for detecting islanding faults in photovoltaic distributed generation systems. The proposed method utilizes a data-driven subspace technique to accurately identify and mitigate islanding events. The verification of the proposed method is conducted through simulation experiments. The islanding detection method presented in this paper does not deliberately introduce any interference sources, thereby minimizing the risk of compromising the power quality of the system. Meanwhile, this technology offers the benefits of rapid detection speed and the absence of any non-detection zone (NDZ).