Temporal and consumer driven cluster analysis for identification of FDI attacks in smart grid

Abstract

The performance and stability of the electric network are improved through smart grid technologies. The smart grid strongly rely on digital communication technologies creates security issues that must be addressed to distribute power efficiently and safely. For billing, load balancing, and energy management, the customer-side smart meter regularly transmits the consumption reading to the system operator. However, dishonest customers created cyberattacks by entering fictitious readings of their electrical consumption in order to steal electricity for financial benefit. The distributed monitoring meter-based network concept is utilized in this paper to closely monitor the losses. A two-stage clustering-based technique has also been suggested for the purpose of detecting theft. The approach can be detect the customer ID having abnormality in consumption pattern. To verify the effectiveness of the suggested strategy, numerical experiments on two datasets of smart meters have been conducted and four different types of attack patterns have been constructed. Simulation results show that the proposed scheme is very effective for detecting various types of attack patterns. In addition, the paper helps to predict how much power and financial loss the company will suffer as a result of multiple attacks.