False Data Injection Attacks (FDIA) in Smart Grid is considered to be the most threatening cyber-physics attack. According to the variety of measurement categories in power system, a new method for false data detection and identification is presented. The main emphasis of our research is that we have equivalent measurement transformation instead of traditional weighted least squares state estimation in the process of SE and identify false data by the residual researching method. In this paper, one FDIA attack case in IEEE 14 bus system is designed by exploiting the MATLAB to test the effectiveness of the algorithm. Using this method the false data can be effectively dealt with.

In allusion to the defect that conventional weight least squares (WLS), when applied to the power system state estimation, generally takes the same weight. It causes large state estimation error. A state estimation method combining optimal weight setting with weighted least squares is proposed. On the basis of weighted least squares state estimation models, through the theoretical analysis of different kinds of voltage measurements, power injection measurements and power flow measurements data, which impact differently on the results, different weights are set corresponding to their measurement sensitivity in order to reduce the estimation errors. Finally, the algorithm is tested on IEEE14 bus system by MATLAB. Compared with the traditional weighted least squares, this method is similar with regard to iteration time. However, the voltage and phase angle state estimation accuracy can be improved by about 0.16 and 4.66 percentage points. Therefore, the algorithm has certain advantages and application value.

Most of methods on malicious data identification are based on the residual in power system applications. Residual error method, which is an effective method to identify a single malicious data can be basically divided into weighted residual error method and normalized residual error method. In this paper the states and measurement estimated value can be calculated firstly by the traditional weighted least squares state estimation algorithm. Then the measurement residual and the objective function value can be also calculated. The algorithm of weighted residual error method is tested on IEEE5 bus system by MATLAB and the analysis on the results of calculation example shows that this method is an effective one which a single malicious data can be effectively dealt with, and it can be applied to malicious data identification. In this paper the largest weighted residues in the case of single malicious data are 8.361 and correspond to real power injection at bus2, which are far above the threshold to improve the efficiency of malicious data identification.