Naive Bayesian islanding detection technique for distributed generation in modern distribution system

In this paper a new islanding detection technique for grid-mode distributed-generation (DG) is proposed. Twenty one features are extracted from measurement of the voltage and frequency at the point of common coupling (PCC) in order to identify islanding occurrence with high accuracy. An IEEE 34-bus system was used in this paper to generate islanding and non-islanding training cases. Then a Support Vector Machine (SVM) was trained using the training cases in order to discriminate islanding and non-islanding cases. In order to test the accuracy of the Naïve Bayesian Classifier, Cross-Validation was used to evaluate the performance of the proposed islanding detection technique. Accuracy of 100% was achieved using the proposed algorithm.