Islanding detection using decision tree approach

Islanding occurs when part of a distribution network is disconnected from the rest of the power system but continues to be energized by a distributed resource (DR). Failure to trip the islanded DR can lead to serious problems for this DR and the connected loads. One method of islanding detection utilizes and combines different system parameters by using data mining technology and a multifunction software model to extract information from these data sets. It also detects the threshold settings for the islanding relays. Using the same methodology, an attempt has been made in this study to simulate the network and obtain the threshold settings by selecting a new set of parameters. Good accuracy was observed when compared to the data mining approach in spite of choosing a different set of parameters. The main advantage is parameters chosen are easily measured and the software tools used are readily available.