Research On PMU Measurement Based Power System Dynamic State Estimation

The accuracy of real-time dynamic state estimation of power system is the basis of sta-bility analysis and control results of power system.The wide area measurement system(WAMS)is based on synchronized phasor measurement technology,can provide real-time,accurate synchronous signals of many nodes in power system,which can be used for solving the dynamic state estimation problem.But the factor that bad data of the wide-area measure-ment system possible problem is for state estimation cannot be ignored,research on bad data identification and repair of PMU measurements method is of great significance.In this paper,the electromechanical transient process of generator is estimated based on PMU measurement system.Firstly,the method of state tracking of the generator and its exci-tation and speed governing system under the fault condition is proposed and the tracking effect has been verified.Then,a bad data identification and correction method based on the Unscented Kalman Filter(UKF)is proposed for the problem of low quality of PMU data which would cause the distortion of state estimation.This method calculated time-varying residual threshold by deriving residual equation.Then,this method can detect the position of bad data by an iterative detection method.For there is bad data,this method rules out this measurement and carries out the estimation again for the purpose of correction.Simulation results have verified the proposed algorithm can effectively restrain the influence of bad data for state estimation.A systematic dynamic state tracking method for power system is proposed in this paper.Firstly,based on the observability theory of dynamic system,considering the load level which has an influence on power system operation point,an optimal PMU configuration method is proposed for dynamic state estimation.The empirical observability Gramian matrix can be used to optimize certain properties of the process state estimates.What's more,the optimal configuration strategy is obtained by the error level determined by the minimum magnitude of the eigenvalue of the empirical observability Gramian matrix.Then,the state of generator and network quantity are separated,and the whole network dynamic state tracking method which is suitable for whole network is proposed by Kalman Filter.Experimental data of dif-ferent systems are used to verify the effectiveness of PMU configuration method and state estimation algorithm.