Consumer phase identification in low-voltage distribution network considering vacant users

Abstract

Accurate consumer phase connectivity in low-voltage distribution network (LVDN) plays a key role in maintaining high reliability of electricity supply and good power quality. However, vacant consumers, a lack of connectivity between feeders and consumers, and measurement error increase the difficulty to identify consumer phase connectivity for the existing data analytics methods. To overcome these hurdles, this paper proposes a novel consumer phase identification (CPI) algorithm based on consumer classification, quadratic programming, and probability distribution. Firstly, a consumer classification method based on voltage characteristics among users is proposed to deal with the vacant user problem. Then, a quadratic programming model based on Nodal Current Law is established to identify consumer phase when the network connectivity information between feeders and consumers is lack. Moreover, in order to improve robustness of the proposed phase identification algorithm on measurement error, a Monte Carlo probability distribution model is developed. The proposed algorithm is applied on a real-world LVDN in Guangdong. The comparison analysis between the proposed method and the Mix Integer Programming (MIP) method, and the impact of the variation rate threshold of correlation coefficient on the identification accuracy are also investigated. The results indicate that the proposed method effectively increases CPI accuracy compared with the MIP method when there are vacant users and measurement error of meters in LVDN.