Query Protection Based On Differential Privacy For Smart Meters

While the popularity of smart meter systems is widespread,the security of smart meter data cannot be ignored.The trusted third party analyzes the energy consumption data through non-intrusive load monitoring technology,and can provide services such as meter charging,power optimization and equipment monitoring for users.However,if a malicious attacker obtains the meter data,it can analyze the user's electricity habits and behavioral activities,which poses a huge threat to the user.The current common privacy-preserving schemes mainly include load-hiding policies based on energy storage and privacypreserving data aggregation.The meter data collected by the smart meter based on the energy storage-based privacy-preserving scheme cannot achieve monitoring of devices,and the calculation operation of the privacy-preserving data aggregation may cause high computational overhead.The privacy-preserving scheme for meter data comprehensively considers the billing function of power supplier and the demand for energy consumption data of third party service,and implements the privacy-preserving query in the meter data management system.The existing privacy-preserving method based on differential privacy sacrifices the accuracy of total power consumption.In order to solve this problem,the scheme uses packet entropy to protect the private information in the data,and also ensures accurate charging of electricity.The scheme includes the request of the third party service for the meter data,and the algorithm of privacy-preserving query is designed.Third-party servers can only obtain query data and perform analysis to provide users with services such as power optimization.The distance score function is used to protect the privacy of queries by using the exponential mechanism.In the proposed privacy-preserving framework for meter data,neither the power supplier nor the third-party service party directly obtains the meter data,but the result data with the differential privacy preserving.Therefore,the proposed solution can protect the privacy of the user without affecting the billing and the user obtaining the service provided by the third party.The data sets used in the experimental analysis are the real energy consumption data sets for household electricity.The experimental results show that the noise adding of privacy protection does not affect the accurate charging of energy consumption data.Moreover,under the same parameter setting of query,the privacy-preserving scheme has a maximum improvement rate of 2.62 times compared with the traditional noise-adding method.