Practical Privacy-Preserving Electricity Theft Detection for Smart Grid

The detection of electricity theft, which focuses on privacy preservation and system security, has been extensively researched in the smart grid. However, existing solutions have not taken into account the enormous communication overhead that will be incurred in practical environments due to the large scale of the smart grid and the vast number of smart meters. Furthermore, most schemes have limited functionalities. There is also a lack of further research on the detection model and period. Therefore, we propose a practical privacy-preserving electricity theft detection scheme. Specifically, we inject the gamma noise into a user's power consumption data to preserve user privacy without adversely affecting the accuracy of detection. Secondly, our approach achieves privacy-preserving dynamic billing and differential privacy for regional power consumption aggregation without requiring any additional operations. Additionally, we propose a novel combination detection model that extracts local and global features of data, and explore the impact of the detection period. Ultimately, we conduct a number of experiments based on real power consumption data and practical devices, and experimental results imply that the proposed scheme can operate on resource-constrained devices with lower communication overhead and better detection model performance compared with sate-of-the-art schemes.