Identifying Topology of Low Voltage Systems via Incorporating Measurement Correlations

Abstract

The management of intelligent power distribution systems is limited due to the lack of observability in low voltage (LV) systems. Ambiguous logical relations among branch units, topological unavailability of consumers increase the difficulty of topology identification (TI). To solve this problem, a three-step framework is proposed via measurement correlations. The first step applies correlations of active power features to identify the up-down and parallel, and phase connection of branch units; the second step specially calculates active power correlations between segments and non-vacant consumers, determines the phase connections, and identifies the connected lines of non-vacant consumers; the third step associates vacant consumers with non-vacant consumers by voltage profiles. Finally, robustness against errors and sample sizes is explored, and the results demonstrate that the proposed algorithm can accurately identify the LV topology based on limited samples of voltage magnitude and active power.