The aim of this paper is to detect the single line to ground fault on the unit generator- transformer. A new ground fault detection scheme
based on the extraction of energy and statistical parameters from wavelet transform based neural network is proposed. The faulty current signals
obtained from a simulation were decomposed through wavelet analysis into various approximations and details. The simulation of the unit generator-transformer was carried out using the Sim-PowerSystem Blockset of MATLAB. The energy and statistical parameters analysis involved measured of the dispersion factors (range and standard deviation) of wavelet coefficients. Regarding the ANN performance, the errors in the SLG�fault detection of ANN were under 1 %. The results indicate that the proposed algorithm was accurate enough in differentiating a single line to ground fault and un-fault for a unit generator-transformer.