Research On Power System Dynamic State Estimation Algorithm Based On Grey Theory

Power system state estimation, which establishes a reliable and perfect real-timedatabase for power system, is an important part of the energy management system, andis also the critical measure to ensure the power system to operate safely andeconomically. Dynamic state estimation can be used to predict and estimate, so theonline features, such as state prediction, security assessment, economic dispatch,prediction and control and so on, can be realized by means of it. It goes without sayingthat the dynamic state estimation is important. The traditional dynamic state estimationis conducted based on the theory of extended Kalman filtering. However, in accordancewith the derivation for the calculation formulation of the extended Kalman filteringalgorithm and the analysis for the work principles, lots of problems are pointed outunder the power system model in practical application. Aiming at the problems that thepredictive model is difficult to establish and the filtering process can not mutatefollowing the system, the thesis introduces the Grey System Theory in detail, and thenthe works of this thesis are summarized as follow:(1) The basic knowledge of grey system theory is presented in detail, and the mainapplications in the field of the load prediction for the power system, electricity demandforecasting and the fault diagnosis are also introduced in this paper.(2) The dynamic state estimation algorithm, which is developed based on the extendedKalman filtering theory, is presented in detail, and then the existed problems in the realapplication are pointed out through the analysis.(3) The correlation analysis for historic state information and the establishment of theassociated sequences for the state are conducted by means of grey relational analysis,and then the associated sequences are modeled through the method of the GM(1,1),obtaining the associated predictive sequences and the predictive values for the states,and the state predictive values are obtained through establishing the objective functionand solving its solution.(4) The traditional filtering algorithm is improved by means of the strong tracking filter,which makes the filtering can track the mutation and the slowly varying for system state,improving the performance of the filtering algorithm.(5) The fault diagnosis algorithms in the power system are introduced briefly, and thenthrough the protection and circuit breaker information, the state vector elements forfaulty component are determined, the assignment rules for the vector elements are set, and the state vectors for faulty component are generated. According to the grey relationanalysis, the faulty components are identified and the protection and circuit breaker arededuced, which enhances the efficiency of the diagnosis algorithm and the anti-jammingcapability.