Phase Identification for Power Supply Network in Commercial Buildings

Understanding of customers and feeder branch associations in low-voltage distribution networks is beneficial to quick fault location, outage response, and asset management. In commercial buildings, the loads are diversified, most of them are three-phase with close electrical distances of service points, which bring more challenges on data measurement, while making the operation and management of the power supply network of commercial areas more difficult. The relatively stable three-phase loads and small difference between phase loads increase the difficulty of phase identification by data driven methodologies. To cope with this problem, this paper proposes a customer phase identification method for commercial building scenarios based on voltage time series analysis. Firstly, voltage time series from smart meters are obtained and pre-processed by interpolation to fill in missing values, then the phase fluctuation difference features are leveraged to extract the segments of voltage time series based on a sliding window scheme for different data collection frequencies, and finally voltage correlation analysis is used to determine phase sequence. The proposed method is tested on measurements from a real-world shopping center under two typical acquisition frequencies and verified the effectiveness and robustness of the method.