Fading Kalman Filter-based Method for Controlled Fault Interruption

Short-circuit current zero-crossing prediction is the prerequisite for accurate breaking of the circuit breaker. In this paper, the fading Kalman filter algorithm is applied to short-circuit current zero-crossing prediction. Considering the feature of the short-circuit current waveform, the prediction covariance matrix is adjusted in real time with the changing fading factor, which improves the convergence speed of the algorithm, thereby improving the zero-crossing prediction accuracy. With the PSCAD simulation model of a simple power system, multiple short-circuit current waveforms are gathered and analyzed with the algorithm proposed, which proves to be effective. With sampling window length of about 5 ms, the prediction accuracy of the first zero-crossing reaches ±0.5 ms. Compared with existing algorithms, the algorithm proposed ensures the prediction accuracy while requiring shorter sampling window length.