A Contactless Method for Unbalanced Loading Detection in Power Distribution Lines by Magnetic Measurements

Reliability of power distribution network is highly desired by end customers. The power utility companies make best efforts to ensure balanced power transfer for single-phase customers. Phase unbalancing is common in congested power distribution systems where a large number of single-phase consumers are served by three-phase transformers. It deteriorates transformer efficiency and results in increased power system losses. As a standard practice, utility providers use portable CTs to periodically measure phase currents to reduce the unbalanced loading of transformers. The approach requires manual intervention and regular inspections. This article proposes a novel noncontact method for the measurement of unbalanced single-phase loading using a single magnetic sensor and processing with a trained neural network (NN). This article employs variation in phase and magnitude of the magnetic field produced by three-phase line conductors to detect and identify the unbalanced lines. The optimal sensing location is determined by using outcomes of coherence between the training and test data. Then, a trained NN uses a pattern recognition algorithm to estimate the unbalanced and the balanced loading. The method is comprehensively tested by numerical simulations and experimentation, which involves various potential configurations of distribution lines (flat, vertical, and delta). The method of estimation of unbalanced loading by magnetic field sensing and pattern recognition method has been experimentally implemented and verified in the laboratory setup. The results provide accurate unbalance phase detection with an error of less than 1%.