On-line dynamic state estimation of power systems

Traditional state estimators based on steady state system model cannot capture the dynamics of power system very well because of the slow updating rate of SCADA systems (several seconds). The emergence of wide-area measurement systems offers new opportunity for developing more effective methods to monitor power system dynamics online. But due to the nonlinearity of state transition and observation equation, linearization and Jacobian Matrix calculation are indispensible in the existing methods of power system state estimation. This makes WAMS' high performance compromised by burdensome calculation and inaccuracy of the state estimation. In order to overcome the drawbacks, this study tries to develop an effective state estimation method without the linearization and Jacobian Matrix calculation. Firstly, unscented transformation is introduced as an effective method to calculate the means and covariances of a random vector undergoing a nonlinear transformation. Secondly, by embedding the unscented transformation in the Kalman Filter process, a new method is developed for power system dynamic state estimation. Finally, some simulation results are presented showing accuracy and easier implementation of the proposed method.