Abstract
This paper presents coherent representations of electric power systems
signals. These representations are obtained by employing adaptive
signal decompositions. They provide a tool to identify structures
composing a signal and constitute an approach to represent a signal
from its identified components. We use the matching pursuits algorithm,
which is a greedy adaptive decomposition, that has the potential
of decomposing a signal into coherent components. The dictionary
employed is composed of damped sinusoids in order to obtain signal
components closely related to power systems phenomena. In addition,
we present an effective method to suppress the pre-echo and post-echo
artifacts that often appear when using the matching pursuits. However,
the use of a dictionary of damped sinusoids alone does not ensure
that the decomposition will be meaningful in physical terms. To overcome
this constraint, we develop a technique leading to efficient coherent
damped-sinusoidal decompositions that are closely related to the
physical phenomena being observed. The effectiveness of the proposed
method for compression of synthetic and natural signals is tested,
obtaining high compression ratios along with high signal-to-noise
ratio.
Users
Please
log in to take part in the discussion (add own reviews or comments).