부산대학교 도서관

An accurate system topology is essential for distribution system monitoring and operation. This paper presents a novel approach to distribution system topology identification (DSTI). The proposed method identifies the topology of the distribution network in real-time by applying Linear Discriminant Analysis (LDA) and Regularized Diagonal Quadratic Discriminant Analysis (RDQDA) to the voltage magnitudes collected by distribution grid sensors. The results show that this method can utilize noisy voltage magnitude readings from load buses to identify distribution system reconfiguration between radial topologies during operation under changing loads in an accurate manner. Furthermore, the approach has shown good results for DSTI utilizing available multi-rate voltage measurements and for reduced-dimensionality input data using principal component analysis. The proposed DSTI approaches have outperformed convolution neural networks under limited training data.