Detection and characterization of oscillatory transients using matching pursuits with a damped sinusoidal dictionary

This paper proposes a reliable method based on matching pursuits for detecting and characterizing power system oscillatory transients. The method compares a portion of the time-domain signal with the damped sinusoidal functions of a redundant dictionary. When the value of correlation between the signal and the function in the dictionary is near to one, it is concluded that an oscillatory disturbance has taken place, and the frequency and damping factor of the function are considered as the parameters of the oscillatory transient. The effects of noise and initial amplitude of transient are numerically examined, and it is found that the method remains accurate when the transient signal is deeply buried in the noise or is of very low initial amplitude. The algorithm does not require any system modeling and the accuracy is not affected by the variations in system conditions, such as network topology or component parameters.