الفهرس 
CHAPTER ONE INTRODUCTION AND PREVIOUS WORKOnly 14 pages are availabe for public view
 |  | from | 244 |  |  |
| | | from | 244 | | |
Abstract
Magnetizing in rush current , MIC , is a natural phenomenon wich occurs during energizing of transformer ,. This high current may reach ten times of full load current and can cause protective relays to mal-operate.
The analysis has been carried out taking into considerations the effect of: residual magnetism , switching on angle , transformer winding connections
Fault current calculations have been carried out for different types of faults on the transformer. The values and waveforms of the short circuit currents have been determined under different types of short circuit for no-load –transformer under load. Effect of the transformer winding connections on the values and waveforms of short circuit currents have been taken into consideration fast Fourier transform , FFT, has been applied , through anew computer program on the waveforms of MIC and fault currents to estimate the harmonic components of each waveform
The discrimination between the MIC and fault currents have been made based on the harmonic component of each waveform fed forward neural network, FFNN, algorithm has been employed for recognition of waveforms of fault current and MIC the computational requirements for the, FFNN, implementation is very large but it is more reliable than the schemes based on detection of only one harmonic. using the propsed computer programs , based on the mathematical models . a large number of fault currents and MIC examples have been obtained , some of these examples have been employed in training and others in testing of the proposed neural network . NN four architectures of FFNN have been trained and tested , using back propagation technique with large sets of examples of inrush and fault current
The test result proved that the purposed FFNN to be adequate and would be useful in building a complete digital protection scheme for disseminations between fault currents and MIC’s