Non-technical loss detection by multi-dimensional outlier analysis on the remote metering data

A multi-dimensional outlier analysis method is presented in this paper for electricity stealing suspect detection (ESSD). A preprocessing method is first presented to resolve the defects in remote metering data. After that, six features are extracted from the customers' daily consumption series, and a cluster-outlier interactive algorithm is applied for solving the ESSD problem. A case study with an estate located in the north of Shanghai, China is provided at the end of the paper. Results show that the presented method can search out the electricity stealers timely in their first months of consumption drop, and is of much higher search rate and much lower misjudgment rate relative to the one-dimensional outlier analysis method.