Search

Frequency shift islanding detection methods have been widely used in inverter-based distributed generations. Two representatives of such methods, Sandia frequency shift (SFS) and reactive current perturbation (RCP) methods, are investigated in this article. The investigation reveals two main issues with these two methods, i.e., parameter selection scheme for positive feedback gain and current static error. The current parameter selection scheme, in particular, is considered unreliable and inconvenient in this article, for it has not considered the weakening effect of PI regulators on positive feedback and the positive feedback gain is actually obtained by trial and error, which makes it difficult to use such methods well in field environments. In view of these issues, this article proposes a solution that contains an improved method to eliminate current static error and a reliable and convenient parameter selection scheme based on system stability analyses. The final simulations and experiments verify the good performance of the improved method and the parameter selection scheme. [ABSTRACT FROM AUTHOR]

This paper first derives a usable formula for frequency shift islanding detection methods on the basis of the parallel R, L, C load and the conclusions in current literature: the angle by which the total output current of the distributed resources (DR) units leads the point of common coupling (PCC) voltage must be conducted to have the same shifting direction as the load admittance angle during the variation of the frequency. According to the formula and multi-DR operation, the scenarios in which the classic frequency shift methods are applied to energy storage converters are analyzed. The results indicate that the setting of the angle by which the energy storage converter current leads the PCC voltage may need to be modified when running state changes. It results in the problems that the classic methods are not applicable for non-unity power factor (non-UPF) control and have to distinguish between generation mode and consumption mode for UPF control. On account of the problems, a coping strategy, i.e., an improved method, is proposed. The analyses indicate that the improved method is applicable in every state. The last simulations and experiments confirm the preceding conclusions. [ABSTRACT FROM AUTHOR]