Smart Grid State Estimation And Performance Analysis

The smart grid combines the traditional electric power technology with the advanced information technology, so that it can ensure the safety, reliability and economical efficiency of power system. In this thesis, the state estimation and its performance in smart grid are investigated. The mathematical model which the state estimation in smart grid will use is formulated. The WLS(Weighted Least Squares) and MAP(Maximum a Posteriori) estimation are discussed. This thesis mainly considers the effect of bad data injected to the grid. The ZZB(Ziv-Zakai Bound) is derived by taking into account the prior distribution of the parameters and relating the estimation problem to a binary hypothesis testing problem. Compared with the BCRB(Bayes Cramer-Rao Bound) and MSE(Mean Square Error), it shows that the ZZB can reflect the influence of bad data, especially the undetected attack, which the BCRB can’t. Then on the basis of the performance analysis, a method to find the optimal PMU(Phasor Measurement Unit) location in smart grid which can optimize the performance of the state estimation under undetected attack is investigated. Aimed at research on the state estimation and performance in smart grid, the thesis will carries out the following work.1. The characteristics of smart grid are presented, and two mathematical models, AC(Alternating Current) model and DC(Direct Current) model are formulated. The estimation method including WLS and MAP are derived for the two models.2. The bad data is defined and classified. Then the smart grid state estimation under bad data attack is discussed and the undetected attack which has serious effect to smart grid is derived. The simulation of state estimation under no attack and bad data attack shows that the effect of bad data to state estimation.3. The performance of smart grid estimation under bad data attack is analyzed. By deriving the MSE and BCRB, the results show that the BCRB can’t reflect the influence of the bad data to state estimation. Therefore, a new bound called ZZB is studied and derived in smart grid. The MSE, BCRB and ZZB are simulated, and the results proved that the ZZB has better performance than BCRB.4. The PMU is introduced, and by defining the PMU location matrix, the mathematical model under bad data attack with PMUs which state estimation will use is formulated. Then the MSE of the state estimation is derived. The optimal PMU location which leads to the minimum MSE is analyzed by establishing the condition optimization problem. The simulation results proved the efficiency and rapidity of this method.