Islanding detection based on impedance estimation using small signal injection

Abstract

This paper presents a new impedance estimation approach for islanding detection (ID). To estimate the impedance seen by the distributed generator (DG), a small signal produced by the DG’s electronic converter is injected into the power system. This signal produces small disturbances in the network that enable the estimation of the impedance without causing major damage to its power quality. In the proposed method, an islanding event is detected if the measured impedance exceeds a pre-defined threshold. This methodology provides better stability against many non-islanding events, avoiding false ID tripping. Additionally, it can identify islanding events even in the case of a perfect power balance condition. Validation was performed using the Hardware-in-the-Loop (HIL) approach. The results showed that the proposed method is superior to an existing active method in the literature for distinguishing islanding and non-islanding events. The methodology can be used for the case where more than two generators are connected at the same access point.