A Hybrid Active Frequency Drift and Pearson Correlation Method for Islanding Detection in Grid-Connected Photovoltaic Systems

Islanding detection remains a critical technical challenge in grid-connected photovoltaic (PV) systems, as delayed or missed detection can compromise both system safety and grid reliability. Conventional Active Frequency Drift (AFD)–based techniques, although effective, suffer from relatively large non-detection zones (NDZs) and prolonged detection times under specific load conditions, particularly for high-quality-factor resonant loads. To address these limitations, this paper proposes a hybrid islanding detection method that integrates AFD-based zero-time-interval perturbation with Pearson correlation analysis. The proposed approach introduces a controlled frequency drift in the inverter output current and evaluates the statistical correlation between the point of common coupling (PCC) frequency and its rate of change. By monitoring the dynamic relationship between these variables, the method reliably distinguishes islanding events from normal grid disturbances with enhanced sensitivity and reduced NDZ. The proposed technique is experimentally validated on a 350 W single-phase grid-connected PV system under IEEE Std. 929 and IEEE Std. 1547 islanding test conditions. Experimental results demonstrate detection times ranging from 0.09 s to 0.18 s across load quality factors of 0.5, 1.0, and 2.5, representing up to a 60% improvement compared to conventional AFD methods. In addition, the NDZ is reduced to below 5%, while inverter current total harmonic distortion remains below 2%, ensuring full compliance with grid interconnection standards. The proposed AFD–Pearson correlation–based method requires no additional hardware or communication infrastructure, making it a cost-effective, scalable, and practical solution for anti-islanding protection in modern grid-connected PV systems.