A High-loss Power Line Theft Detection Method Based on Segmented Dynamic Time Warping Distance

Most of the existing electricity theft detection approach of distribution lines identifies anomaly users with large volume of electricity usage data. Since users’ electricity usage data and loss of associated distribution lines are highly correlated, this could be analysed to identify the dishonest user of electricity theft. A segmented dynamic time bending distance based electricity theft detection approach of high-loss line is proposed in this paper. Firstly, the heuristic segmentation algorithm is used to transform the time series of electricity usage of each user and line loss to implement feature extraction and data reduction. Thereafter, dynamic time bending distance is employed to find out the user whose electricity usage is most similar to that of the line loss. The user whose electricity consumption has most similarity to that of line loss with the same fluctuation direction is assigned as the suspected user of electricity theft. Numerical simulation of real world metering data of high-loss lines suggest that the proposed approach has better accuracy and lower false positive rate as compared to the comparison method.