Bad Data Detection And Identification In Power System

Power system state estimation is one of the most important parts of modern energy management system (EMS) ,its measurement contains bad data ,besides normal noise.Bad data makes the result of estimation inaccurate,so bad data detection and indentification is the necessary technical measures in power system state estimation.According to their stage of the process in power system state estimation, bad data detection and identification methods can be divided into three categories, such as before estimation, in estimation and after estimation,and most of them belong to the third category—after estimation ,their disadvantages are a large amount of calculation, long computing time and miscarriage of justice or omission due to residual mask and residual transfer phenomenon.In this paper, two kinds of bad data detection and identification methods are applied .They belong to the category of before estimation and use correlation coefficient of measurement variations and innovation graph,and improvments were added to the latter.The former makes use of correlation coefficient of measurement variations to detect and identify bad data, through analyzing the correlativity of measurement and the correlativity of residual. Simulation based on IEEE9-bus network results show that this algorithm is able to distinguish bad data from sudden change data, meanwhile it can identify multi bad data once with low failing ratio.The latter is based on dynamic and static theory, it translates the difference between the recent measurement and forecasted values (innovation vector) into a equivalent variable of another abstract domain, using the basic theory of circuit, for detecting topology error and bad data and identification. This method is of simple model,speediness and nicety, and its efficacy was verified by the exampleof IEEE14-bus network.