State Estimation In Three-phase Unbalanced Distribution System With False Data Injection Attacks

State estimation of distribution system estimates the real-time system operating state based on sparse measurement data,and provides data support for important decisions such as load distribution.With the intelligent development of the power grid,the coupling of information system and physical system of the distribution system has gradually improved its situation awareness,it has also provided more entrances for injecting false data injection attackers at the same time.False data injection attacks could severely damage the reliability of distribution system by disturbing state estimation result and misleading important decisions.Therefore,this dissertation focuses on the methods for state estimation in three-phase unbalanced distribution system under consideration of false data injection attacks.The specific works are as follows:1)Research background and significance of the subject are firstly introduced.Domestic and international researches are reviewed from two aspects including the detection methods of false data injection attacks and the optimal configuration of defense sources.This dissertation divided the existing methods into categories and analyzed the shortcomings of existing methods,introduced research ideas.2)According to the characteristics of three-phase unbalanced distribution system and multi-source measurement data,combined measurement transformation technology with maximum normalized residual test,proposed a linear state estimation model for three-phase unbalanced distribution network.A false data injection attack vector constructed based on the vulnerability of bad data detection,and the simulation example verified the validity of the threephase unbalanced distribution system state estimation and the feasibility of the false data injection attacks.3)With the consideration of the large scale and low computational efficiency of the distribution network,research on false data injection attack detection algorithms based on distributed state estimation.Combined the node-overlapping partition method and the subregion data interaction method,a distributed state estimation framework is established.Based on the Kullback-Leibler distance,a distributed detection algorithm for false data injection attacks is proposed.The simulation example verified the reliability of the results of the distributed state estimation and the effectiveness of the distributed detection method.4)Aiming at the security risks of cyber-physical distribution system,the attack path diagram is constructed with pseudo measurement and real-time measurement as the attack target.Combined the false data injection attack success probability and protection cost to establish the security risk quantification mechanism for cyber-physical distribution system.Established an optimal allocation model of defense resources for the area vector measurement device.The simulation example verified the feasibility and effectiveness of the proposed security risk assessment system and defense resource optimal allocation method.