Research On Privacy-preserving Methods Based On Data Perturbation For Smart Meter

In recent years,with the construction of smart power grids and smart meter,the development of social intelligent electricity network is promoted and the labor structure optimized,as well as the power resource allocation improved.While enjoying the improvement of quality of life brought by electricity,people can experience various intelligent services based on power data.However,fine-grained data from smart meter not only bring intelligent services,but also may leak private information.By analyzing the data of smart meters,attackers can obtain information such as users’ behavior patterns,habits,domestic economy and personal preferences,which can causes serious disclosure of personal privacy.As for users,they are unwilling to provide personal data to service providers for the considerations of protecting their own privacy,which not only affects the advancement of smart grid construction,also fails to exploit the potential value hidden in big data of electric power.Therefore,it is of great significance to study the privacy-preserving methods for smart meters.Data perturbation is a common privacy-preserving method for smart meter,It mainly includes three algorithms which are random perturbation,charging battery and differential privacy.To a certain extent,the current privacy-preserving methods for smart meters based on data perturbation can protect users’ privacy,but there are still some problems.On the one hand,the existing algorithms do not consider the particularity of privacy protection for smart meters,and their protection for data pattern is inadequate.On the other hand,the current privacy protection algorithms pay more attention to data security while ignoring data availability in different aspects.This paper focuses on the key issue of privacy-preserving method based on data perturbation for smart meters and proposes a temporal perturbation based privacy-preserving scheme,which has good performance in security and availability of data.The main contributions of this paper are listed as follows:(1)Studying the principle of privacy disclosure in smart meter,we point out that the essence of smart meter privacy protection is the protection of mode instead of the traditional protection of single data.The traditional privacy-preserving method which focus on single data is not applicable to smart meter data.And we propose the mechanism of mode protection for smart meters data.In order to satisfy the security requirement of privacy protection,we need to destroy the original waveform recordedby the meters,thus fuzzy users’ behavior pattern,and achieving the purpose of protecting users’ privacy.(2)On the basis of random disturbance and charging battery,we propose a privacy-preserving algorithm for smart meters based on time disturbance,which transfers the idea of random disturbance in numerical domain to time dimension.By increasing the data movement on the time axis,the ability of resisting filter is improved.Furthermore,the ability of the charging battery in pattern perturbation is also improved,which bring out more efficient waveform destruction and achieve the purpose on protection for power pattern.We evaluate the availability of smart meter data in power metering and aggregated data mining respectively by cumulative time error of single-user and data aggregation error of multi-user.On the security,firstly,the destructive effect of power waveform is quantified by root mean square error.Also we use non-intrusive load monitoring algorithm as an attack method to evaluate the security of our algorithm.The experimental results show that our method has good performance in privacy security and data availability,comparing with random perturbation and charging battery.