Abstract
The urban distribution network contains a significant number of multi- or three-terminal connections. Fault location in such networks is severely constrained by the lack of necessary conditions for deploying measuring instruments at intermediate connection points and the presence of erroneous or missing distribution line characteristics. This study designs a fault determination time index by analyzing the changes in Micro-phasor measurement unit (µPMU) measurement data at end nodes when an unbalanced fault occurs in the distribution line. A fault localization model incorporating frequency parameters is proposed, considering the influence of source load variations and the regular fluctuations of the fundamental frequency, which affect the characteristics of the distribution lines. To enhance the model's accuracy and efficiency, a solution approach combining Simulated Annealing Algorithm and trust region methods is suggested, addressing the impact of the initial values on the fault localization model calculation process. A three-terminal distribution line model is constructed in Matlab, and various failure scenarios are simulated. Using these frequency values significantly improved the accuracy of the fault location model's estimation results, achieving accuracy more than three times higher than models that do not consider frequency. The fault distance estimation error is reduced to less than 50 m. And, the model's applicable fault scenario is increased to 2000Ω. The results demonstrate that the proposed technique significantly improves the calculation efficiency and accuracy of the fault location model, providing a robust solution for fault location and line parameter estimation in distribution networks.