Research On Islanding Detection Techniques Based On Grid-connected Distributed Generation Systems

Islanding detection is an indispensable feature for the grid-connected DG system tooperate safely and trustfully. Now the islanding detection mainly concludes remote sensingdetection method and local detection method, but each specific method has certainrestriction condition and the drawbacks of detection. To extract the voltage and frequency atpoint of common coupling (PCC), most of schemes still adopt one of voltage-over-zerodetection method, inverse Park transformation-DPLL method and FFT spectrum analysismethod. But all these methods can not satisfy the requirements of speed, accuracy and thereal-time performance of islanding detection.In this paper, a PCC voltage and the frequency extraction method is proposed, which isthe "low-pass filter+adaptive notch filter (ANF)" method. The low-pass filter is used to filterout the harmonic components of the PCC voltage, and then using the ANF to extract thevoltage amplitude and frequency. The “islanding detection+DSP+second-order low-passfilter+oscilloscope” hardware system is built to verify the proposed voltage and frequencyextraction method by experiment. The simulation and experiment results show theeffectiveness of the proposed method: when the PCC distorted voltage contains harmonics,the proposed method can extract the fundamental voltage amplitude and frequency inreal-time, and only needs two power frequency cycles (0.04s), which is fast and precise andcan fully meet the requirements of the islanding detection standards.