الفهرس Only 14 pages are availabe for public view
 |  | from | 173 |  |  |
| | | from | 173 | | |
Abstract
Power transformers are a very expensive class of electrical devices. They represent a very important link between generating utilities with transmission systems and between transmission systems with load centers. Internal windings faults constitute a large percentage of transformer failures, yet the available PSCAD/EMTDC software doesn’t simulate such type of faults. The simulation of this type of faults is essential to generate data that is needed in protection studies to test and evaluate the performance of digital relays. The first part of this thesis presents a transformer windings model which proves to be an efficient way to represent the transformer internal windings faults. The proposed model uses the steady state resistance and inductance matrices of the transformer, which can be obtained from excitation and short circuit test data, to calculate the matrices representing the transformer during turn to ground or turn to turn fault utilizing the principles of consistency, proportionality and the leakage factor of the windings. The calculated matrices are used in writing the circuit equations that represent the faulty conditions of the transformer. The proposed model is implemented via a MA TLAB program that generates the current and e.m.f samples during the faulty conditions, the simulation results proved to be accurate announcing the validity of the model.
The Differential protection is considered to be the main protection of many power system components; however, it suffers a main drawback when applied to power transformers, as false tripping may occur due to differential currents caused by healthy cases like transformer magnetizing inrush currents, or by other faulty cases which aren’t cleared by differential protection like over excitation and external faults.
The second part of this thesis introduces the wavelet transform as a new and efficient tool for signal analysis and applying the wavelet transform to discriminate between differential currents due to internal faults and differential currents due to magnetizing inrush current. The discrete wavelet transform (OWT) decomposes the differential current signals using the mother wavelet ”db2” into five levels, and calculates the spectral energies for these levels, among which the approximation, fifth, and fourth details are selected to study their variation. This variation is used to calculate a characteristic quantity and to set operating thresholds that are used to discriminate between faulty and healthy cases. This basis of discrimination is then integrated into the logic of the differential relay. Simulated faulty and healthy cases are introduced to the relay, it has shown high reliability, selectivity and fast response.
The faulty and healthy cases are studied using the PSCAD/EMTDC software, and proposed MA TLAB model for internal windings faults. The simulated data is transferred via an interface unit from PSCAD to the MA TLAB environment where all data processing, wavelet transform, and relay logic programs are performed. The simulated data is used to prove the validity of the proposed relaying algorithm.