Upsampling Impedance Data in Existing Network Datasets Considering Likely Errors

Four-wire low-voltage distribution networks invariably have unbalanced impedance, and overhead lines and cables of different configurations exist. Many network datasets represent impedance in sequence components, which inherently assumes balanced network impedance. This assumption impacts impedance data quality and compromises the accuracy of power flow studies. To improve data quality of low-voltage networks, this paper proposes a methodology to up-sample impedance data given in sequence components, through maximum likelihood estimation with noisy data and mapping to standard conductor and geometry types. Bad sequence component data are identified and removed before the up-sampling process. A proof-of-concept study is conducted to determine whether the diagonal sequence components belong to an overhead line or a cable. Presented results include an up-sampled version of a publicly available Australian low-voltage network model, which is also made available.