Research On Method Of Grid Equipment Parameter Identification And Estimation Based On Multi-period Measurement

The accuracy of grid equipment parameters is the base of the power system analyzing and computing softwares. For various reasons, there are always some errors for line and transformer parameters, thus affecting the reliability of on-line or off-line calculation program. Aiming at this problem, the paper carries on research to the method of parameter identification and estimation, based on multi-period PMU or SCADA data of single equipment. The contents presented in the paper are as follows:Firstly, considering fully the Normal Distribution characteristic of random measurement error, the paper proposes two mean identification methods for parameter errors respectively based on multi-period PMU and SCADA data of single equipment. First, based on the equivalent circuit model of line, double winding transformer and three-winding transformer, and their PMU and SCADA measurement , the two methods respectively create comprehensive normalized value residual index of single equipment at the same period, which is to comprehensively reflect the influence of measurement and parameter errors on residual. Then convergence condition of variance coefficient is used to get the mean absolute value of algebraic sum of multi-period comprehensive residual (T index) ,in order to extrude the influence of parameter errors on comprehensive residual mean value. At last, according to the mean value of comprehensive residual, the parameter errors of the equipment can be identified. This paper analyzes the problem of the identification method in the present literature, which uses the style of the square sum mean value of comprehensive residual with parameter errors. At the same time, the effectiveness of the paper method is verified by simulation analysis.Secondly, considering the problem of randomly changing parameter estimated values with the impact of measurement errors, this paper respectively presents two robust estimation methods of transformer reactance based on PMU and SCADA multi-period measurement information of three-winding transformer at three sides. First, regarded the transformer three side's branch current phasor, neutral point voltage and reactance as state variables, the paper establishes three-winding transformer multi-period PMU or SCADA measurement equations and corresponding extended least squares reactance estimation model, and uses Gauss-Newton method to solve it. Second, regarded the least squares estimated values of transformer reactance as the sample, regarded the variance coefficient threshold of random sample as the convergence criterion, get the average value of the transformer reactance's least squares estimated value by iterative calculation,and make it as the final parameter estimated values. Because of using the average estimation method and choosing the proper average samples, the proposed method has outstanding robust superiority. With the different load levels of typical 500kV three-winding transformer, this paper analyzes the random character of its least squares parameter estimated values, and also verifies the correctness of the paper model and method.