A new Kalman filter approach to digital relaying

Analog prefiltering in computer relaying is a low-cost approach for reducing the magnitude of the noise associated with fault-induced transients. However, analog prefilters are themselves a source of transient distortion over the signals. It is shown how to include in the Kalman filter state space the analog prefilters´ transients in a simple way. Also, it is shown how the coloring effect that the prefilters and the network itself have upon the noise can also be handled by slightly augmenting the system model. The new Kalman filter algorithm is used to support a symmetrical components based distance estimator with moderate sampling rate. The approach is extensively simulated on EMTP data, showing that accurate results are always reachable between 1/4 and 1/2 cycle of the fundamental sinusoids, even for very noisy fault observations