Phase Identification Method for Low-Voltage Distribution Network Users Based on Optimization Model

The accuracy of the phase connection relationship among low-voltage (LV) users significantly affects the optimization of the network operating status. Based on the electrical quantity data collected by smart meters, a method for phase identification of users in a low-voltage distribution network (L VDN) is proposed. First, a three-phase LinDistFlow model is established for the L VDN to obtain the coefficient between the square of the user node voltage and the square of the phase voltage on the secondary side of the transformer. This coefficient is characterized by a 0–1 distribution, which provides principle support for phase identification. Then, the phase identification problem is transformed into a 0–1 optimization problem and the Mixed-Integer Linear Programming (MILP) model is constructed. Finally, based on the historical data of voltage, active power, and reactive power collected by the smart meters at the LV side of the transformer and at each user within the LVDN, phase identification is achieved by solving the MILP model. The simulation example results show that the method has 100% identification accuracy and strong stability, which can effectively resist the interference of smart meter sampling error. At the same time, this method uses only the voltage and power data collected by the smart meters on the secondary side of the transformer and the user nodes for phase identification. It is of high practical engineering value because it works even when the LVDN topology, line parameters, and phase of all users are unknown.