Adaptive estimation of power system frequency deviation and its rate of change for calculating sudden power system overloads

A novel Kalman filtering-based technique is presented for estimating power system frequency deviation and its average rate of change during emergency operating conditions that may require load shedding. This method obtains the optimal estimate of the power system frequency deviation from noisy voltage samples and the best estimate of the mean system frequency deviation and its rate of change while accounting for low-frequency synchronizing oscillations which occur during large disturbances. The proposed technique is a two-stage algorithm which uses an adaptive extended Kalman filter in series with an adaptive linear Kalman filter. The extended Kalman filter calculates the frequency deviation, magnitude, and phase angle of the voltage phasor, which may change during the time period covered by the estimation window. Both the measurement noise variance and the system noise covariance associated with the voltage samples are calculated online. The instantaneous frequency deviation is used as the input to a linear Kalman filter, which models the frequency deviation as a random walk plus a random ramp process. The estimated average rate of frequency decay is represented by the slope of the random ramp. Results for both single and multiple measurements are reported