Harmonic Signature-Based One-Class Classifier for Islanding Detection in Microgrids

This article presents a new passive islanding detection technique in MGs that uses locally measured voltage signals at the PoC of DERs. The proposed method distinguishes events from normal/non-islanding conditions by utilizing superimposed harmonic spectra extracted through full-cycle discrete Fourier transform. Our solution utilizes machine-learning-based one-class classifier to define and adjust thresholds for full spectra. Unlike other methods, our approach does not require data synchronization or communication infrastructure, nor it suffer common errors often arise current transformers. Moreover, design is compatible with distributed decentralized control strategies, as relies solely on local measurements PoC. Another advantage this its low sampling frequency requirement, range 1 kHz, making cost-effective implementable most existing systems. In comprehensive evaluation typical MG test system included synchronous inverter-based DERs, scheme demonstrated exceptional performance. Specifically, was able detect 99.06% different within training range, time just 10 21 ms. Additionally, remained 100% stable during various normal conditions, short-circuit faults, load changes, capacitor switching, changes.