Summary
Power systems signal compression is an important task considering the nature of electrical disturbances, which must be analyzed offline in order to extract some useful information. This paper presents a comparative study about greedy algorithms to obtain sparse representation of signals applied to power systems signal compression. The algorithms will be compared according to two parameters: the quality of the compressed signal, in terms of its correlation coefficient related to the original signal, and the number of elements in the representation, which is related to the compression ratio. In addition, a computational complexity, in terms of required flops, will be performed, aiming at identifying which algorithm is most suitable to run in real time. Finally, a comparison with the wavelet transform, using a database of real power system signals, will be shown.