Page 7 - CE_Industral_Journal_Jan2008

A Novel Approach for Ground Fault Detection

considered non stationary, temporally varying, and
Acquisition
of various burst duration. Even though the current
signals do resemble the third harmonic component
of the current, other harmonic components as well
Data Filtering as non-harmonic components can play a vital role
in the high impedance ground fault detection. One
challenge is to develop a data model that acknowl-
edges that high impedance ground faults could take
2nd Order Neural
Statistical Wavelet Network place at any time within the observation window of
the current signal and could be delayed randomly
and attenuated substantially depending on the
fault location away from the measuring station. The
exemplary model is motivated by previous research,
Decision Logic actual experimental observations in the laboratory,
and what traditionally represents an accurate depic-
tion of a non stationary signal with time dependent
Detection Decision spectrum.
Figure 2. Composite high impedance ground fault detection system. The high impedance ground fault detection problem
addressed in this paper is formulated as such:
Signal Model
Hypothesis H0 : r(t) = s(t) + n(t)
(1)
The high impedance ground fault detection algorithms
reported in this paper are based on using current sig- Hypothesis H1 : r(t) = s(t) + n(t) + f(t)
natures in all of the 3-phases and/or ground which are
(2)

Industry Journal 6