Research And Implementation Of Privacy Protection Technology In Smart Grid

With the continuous improvement of urbanization,smart grids have attracted more and more attention.It combines traditional grids and information technology,enabling bidirectional transmission of information between users and power companies.In this scenario,users upload fine-grained power consumption data to companies,and companies collect and analyze data to achieve reasonable power allocation.However,collection and analysis of fine-grained information poses the risk of leaking users’ privacy,which may cause users’ property loss and even endanger personal safety.This dissertation studies the privacy protection technology in smart grid from the two stages of power data collection and power data analysis.The main contributions are as follows:(1)The study proposes a privacy protection user billing scheme with a smart grid architecture where there is no trusted third-party.As far as the data collection stage is concerned,the existing privacy protection technologies are mostly based on the smart grid architecture where trusted third-party exists.In addition,in order to protect the privacy of power consumption data,existing schemes mostly break the association between users and data,so that power companies can only obtain aggregate information,thus introducing the billing problem.In view of these problems,this dissertation proposes a privacy protection user billing scheme.The scheme supports real-time electricity tariff billing,user dynamic joining and withdrawal,and bill verification in a smart grid architecture where there is no trusted third-party,while protecting users’privacy.(2)The study proposes a clustering algorithm for electricity consumption data in smart grid that satisfies differential privacy protection.Existing data mining analysis researches have been weaker in the privacy protection of power data.Based on the k-means algorithm,this dissertation proposes a power consumption data clustering algorithm in smart grid that satisfies differential privacy protection.Firstly,the algorithm uses density optimization and combines with the exponential mechanism to optimize the selection of the initial center point,while taking privacy protection into account.Secondly,for the dynamic joining of power users,the algorithm introduces two evaluation indexes of intra-class similarity and inter-class similarity,and adjusts local data through splitting and merging,thereby reducing computing overhead and improving clustering efficiency.Analysis shows that the algorithm meetsε-differential privacy protection.Not only that,the simulation result also shows that the algorithm is reasonable in the selection of the initial center point,and the clustering results are available.(3)The study designed and implemented a smart grid power load clustering system based on differential privacy protection.Based on the above-mentioned clustering algorithm,this system can perform statistical analysis in different dimensions,obtain user behavior patterns in various regions,determine regional user types,and formulate corresponding power value-added services based on power consumption data provided by data providers,so that power companies can better serve users.