Performance evaluation of active islanding-detection algorithms in distributed-generation photovoltaic systems: two inverters case

Abstract

Grid-connected photovoltaic (PV) inverters employ an islanding-detection functionality in order to determine the status of the electrical grid. In fact, the inverter must be stopped once the islanding operating mode is detected according to standards and grid-code limits. Diverse islanding-detection algorithms have been proposed in literature to cope with this safety requirement. Among them, active methods based on the deliberate perturbation of the inverter behavior can minimize the so-called nondetection zone, which is a range of conditions in which the inverter does not recognize that it is operating in an undesired island. In most cases, the performances of these methods have been analyzed considering a highly dispersed generation scheme, where only one distributed-generation power system is connected to the local electrical power system (EPS). However, in some studies, it has been highlighted that if two or more PV inverters are connected to the same local EPS, their anti-islanding algorithms do not behave ideally and can fail in detecting the islanding condition. However, there is no systematic study that has investigated the overall capability of different anti-islanding methods employed on several inverters connected to the same EPS to detect islanding condition. This paper is a first attempt to carry out a systematic study of the performances of the most common active detection methods in a case of two inverters connected to the same EPS. In order to evaluate the global capability of the two systems to detect islanding condition, a new performance index is introduced and applied also to the case when the two inverters employ different anti-islanding algorithms.