Electricity theft Detection with Borderline-SMOTE and WDCBL in the Smart Grid

Abstract

The non-technical loss caused by electricity theft on the user side not only increases the operation
cost of the smart grid but also causes some harm to the electricity system. At present, the existing
electricity theft detection model ignores the time-series correlation of user power consumption data,
and the distribution of positive and negative samples in the relevant data sets is uneven. Aiming at
the existing problems, we propose an electricity theft detection model based on Borderline-SMOTE
and WDCBL(wide and deep CBL) networks. The WDCBL model consists of wide components
and deep CBL components. The wide part uses a fully connected layer to extract one-dimensional
features in user electricity consumption data. The model introduces the hybrid network model of
a one-dimensional convolutional neural network(CNN) and bidirectional long-term and short-term
memory network(BiLSTM) into the deep part, which can learn both spatial and temporal features in
user electricity consumption data. Finally, we conducted a comparative experiment on the user electricity
consumption data published by SGCC. The results show that our model is superior to others
in the comprehensive performance.