Abstract :
In this paper, a new method of real-time dynamic state estimation is proposed, in which the model parameters such as state transition matrix, model error and measurement error covariance matrices are all identified on-line. In this way, the degenerate estimated values caused by inaccurate state transition equations are removed, and the estimated precision is high. The method is employed in different scenarios while the estimation performance can be maintained effectively. This approach has been tested on IEEE 5-bus test system under various operating conditions, where normal operation, bad measurement, sudden load change / drastic generation variation are all investigated. The performance indices under different test cases are all evaluated. Test results support the feasibility of the proposed method for dynamic state estimation applications.
Keywords :
covariance matrices; load forecasting; measurement errors; power system state estimation; IEEE 5-bus test system; bad measurement; covariance matrices; drastic generation variation; dynamic state estimation; load forecasting; measurement error; power system; state transition matrix; sudden load change; Covariance matrix; Energy measurement; Gain measurement; Power measurement; Power system dynamics; Power system measurements; Power systems; Smoothing methods; State estimation; Time measurement; extended Kalman filtering; power system; realtime state estimation; weighted least squares(WLS);
