Research On Attack And Defense Of Power System FDIA Based On Data Cleaning And Fair Weight Function

In recent years,False Data Injection Attacks(FDIA)has entered the researcher’s field of vision.FDIA as a new type of damage mode that uses traditional bad data detection vulnerability to launch network attack on power network,can seriously affect the core function of SCADA system--State estimation,causing great damage to the security and stable operation of the power system.First of all,this paper introduces the basic principle and mathematic model of state estimation in detail,combined with the principle of traditional bad data detection,expounds the theoretical basis and attack mode of FDIA.The experiment of the DC power flow model is conducted to attack the specific targets and random targets.The results show that no matter what kind of attack,it can avoid traditional bad data detection and cause damage to the power grid.Secondly,in view of the characteristics of FDIA and the existing research on the theory of Resistance,this paper proposes a defense method based on the combination of data cleaning and Fair weight function.Raw data for state estimation before entering the SCADA system,will be divided into line transmission power,load node injection power,generator node injection power 3 types of measurement data,each type of measurement data using the corresponding cleaning rules for cleaning,cleaning out unqualified data after the input system for State estimation.Due to gross error or dirty cleaning,it may also cause problems in the estimation results of the system.So the least squares algorithm based on the Fair weight function is used to compute.According to the change of the residual variation,and the final result is further corrected.Finally,through the simulation calculation on IEEE 5-node Test system and IEEE14-node test system,the validity of this method is validated by analyzing a large number of data.