Privacy Preserving User Energy Consumption Profiling: From Theory to Application

The smart grid benefits and suffers from smart meter data. Proper use of massive data can improve energy services but may raise privacy concerns. For example, user energy consumption profiling, a classic method, can identify energy consumption patterns based on the collected load profiles from users. Thus, the privacy of these individual load profiles needs to be protected. However, most of the existing works focus on data transmission and calculation privacy, and often require additional computation, communication, or platform construction costs. In contrast, noise-injection-based data source privacy-protecting works can avoid such additional costs and provide theoretical differential privacy (DP) guarantee. This paper theoretically analyzes noise-injection-based user profiling mechanisms in terms of both privacy protection and accuracy. Specifically, we establish the privacy-accuracy trade-off. We then propose an optimal user energy consumption pattern estimation method for heterogeneous noise-injection-based data. Finally, we design a valid information ratio-based pricing scheme for noisy data that is independent of downstream tasks and easy to implement. Numerical studies based on field data confirm the effectiveness of our theoretical results.