Anti-Islanding Methods for Inverter-Based Distributed Generation Systems

Unintentional islanding conditions of grid-connected distributed generation (DG) systems pose significant safety threats to power line personnel and power system equipment. IEEE standard 1547(TM)-2003 requires a fast shut down of the grid-connected DG system once it experiences an islanding condition from the main power system. Although passive islanding detection techniques have no negative impact on DG performance, they possess a shortcoming characterized by a large non-detection zone (NDZ). Active anti-islanding techniques, on the other hand, effectively reduce the NDZ, but the perturbations they introduce degrade the performance of the DGs.;New islanding detection and protection methods have been developed and presented in this thesis. The techniques aim to realize highly sensitive and robust islanding detection. A recessive allele principle is firstly proposed to detect islanding. The embedded recessive allele filter, hibernating in grid-connected operation, will express itself distinctly when islanding occurs. During islanding operation, the behavior of the recessive allele is detected by a proportional power spectral density (PPSD) estimator, a proportional energy (PE) estimator and a covariance estimator of the period of the voltage at the point of common coupling (PCC). An accelerated logic phase shift (ALPS) algorithm and an accelerated reactive power shift (ARPS) algorithm are also proposed as additional robust active anti-islanding methods for single-phase and three-phase inverter systems. A hybrid anti-islanding algorithm based on the combination of PPSD or covariance estimator as the islanding indicator and the ALPS or ARPS as the islanding eliminator is shown to possess excellent anti-islanding performance.