Measuring Trustworthiness of Smart Meters Leveraging Household Energy Consumption Profile

Smart grids use Internet of Things (IoT) devices and play a vital role in promoting a sustainable environment by balancing energies received from different sources. Like other IoT-based systems, cybersecurity of smart meters is an important issue, as data breaches can cause imbalanced load distribution and increased electricity cost. Therefore, cybersecurity protection for smart meters can advance sustainable energy systems. A notable recent approach assessing smart meters' trustworthiness assumes the difference between arithmetic and harmonic means of meter readings as invariant, which limits its applicability and efficacy in different geographical locations with varying electricity consumption needs. The authors introduce an adaptive profile-based trustworthiness measure of smart meters without any assumption on energy consumption data. The household energy profile is generated using historical data and updated with new observations. Trustworthiness of new observations, derived from energy profile changes due to attacks and the distance from its profile, is fused using Dempster-Shafer theory. Such profile- and distance-based evidence makes trust assessment more effective and robust. Results derived from the London smart meter dataset show that the proposed model outperforms the aforementioned approach with average improvement of F1-score (29%), false positive rate (28%), and false negative rate (37%), and better detection rate (98% versus 80%).