Data Driven Energy Theft Localization in a Distribution Network

Energy theft unsettles the regular operation of utility planners and can make the grid more vulnerable. The customers behind the meter can also undergo both financial and security risks for any energy theft happening in the electrical network. This paper presents a straightforward approach to detect the electricity theft event at nodes in the distribution network. The proposed methodology utilizes 15-minute voltage magnitude profiles at all the nodes in the network and investigates the electricity theft node by using the statistical properties of the available data. A test network with 59 nodes is employed to test this algorithm. Results show that this approach can provide a reasonable accuracy to localize the node of the electricity theft event.