Intelligent Island detection method of DC microgrid based on Adaboost algorithm

Abstract

Aiming at the problems of large detection blind areas and difficult threshold settings in traditional passive islanding detection methods, this paper proposes an intelligent passive islanding detection method based on the adaptive boosting (Adaboost) algorithm. This method first uses feature screening technology to obtain the key feature electrical quantities of DC microgrid islands and form a sample set, and then rely on the Adaboost algorithm to generate a high-precision island classification model to realize the detection and classification of DC microgrid operation status. It has the advantages of automatic setting threshold, small detection blind area, and high accuracy. In addition, aiming at the problem that the classification results of the intelligent islanding detection algorithm based on machine learning cannot be directly used as the islanding discrimination results, an islanding criterion method based on the sliding window algorithm is added on the basis of the classification algorithm. This method realizes the accurate detection of the DC microgrid islanding operation state by the secondary analysis and judgment of the classification results of the islanding classification model. Finally, a DC microgrid model is built to simulate and verify the proposed method. The results show that the method can quickly and accurately complete the islanding detection of the DC microgrid.