Theft Detection Unit For Photo-Votaic Generation in Smart Grid Networks

Abstract

While the increased connectivity of the power grid has allowed for the automation of its functionality, it has also led to a heightened vulnerability to cyber threats, putting the whole power system security at risk of energy theft through the manipulation of data. In addition, the introduction of the smart grid allows customers to have their own power-generating units, which are usually photovoltaic (PV) panels. With two-way communication under the smart grid paradigm, customers’ local generation can be measured by smart meters and reported to the utility, which in turn pays customers for their generated electricity. Manipulating smart meters to report false generated electricity is a growing concern that can jeopardize a utility’s revenues. Thus, the objective of this work is to design and build an intelligent theft detector unit for PV injection (TDUPV) that detects suspicious data flow from customers’ solar smart meters to the back-end system within the utility. This topic contributes to the theft detection research community as it considers the injection of PV panels, which had not been considered in any previous research work. The detector is based on a regression tree model that utilizes weather information and customers’ PV injections to predict the honesty of the injected power from customers’ PV panels reported by the solar smart meters, assuming a data flow manipulated by cyberattacks. The mechanism of detection is based on the probability density function (PDF) of the error between the actual and predicted values. The performance of the TDUPV was evaluated by testing several case studies under different theft scenarios and shows the effectiveness of the proposed unit.